Elevated Cell Wall Chitin in Candida albicans Confers Echinocandin Resistance In Vivo

Abstract

Candida albicans cells with increased cell wall chitin have reduced echinocandin susceptibility in vitro. The aim of this study was to investigate whether C. albicans cells with elevated chitin levels have reduced echinocandin susceptibility in vivo. BALB/c mice were infected with C. albicans cells with normal chitin levels and compared to mice infected with high-chitin cells. Caspofungin therapy was initiated at 24 h postinfection. Mice infected with chitin-normal cells were successfully treated with caspofungin, as indicated by reduced kidney fungal burdens, reduced weight loss, and decreased C. albicans density in kidney lesions. In contrast, mice infected with high-chitin C. albicans cells were less susceptible to caspofungin, as they had higher kidney fungal burdens and greater weight loss during early infection. Cells recovered from mouse kidneys at 24 h postinfection with high-chitin cells had 1.6-fold higher chitin levels than cells from mice infected with chitin-normal cells and maintained a significantly reduced susceptibility to caspofungin when tested in vitro. At 48 h postinfection, caspofungin treatment induced a further increase in chitin content of C. albicans cells harvested from kidneys compared to saline treatment. Some of the recovered clones had acquired, at a low frequency, a point mutation in FKS1 resulting in a S645Y amino acid substitution, a mutation known to confer echinocandin resistance. This occurred even in cells that had not been exposed to caspofungin. Our results suggest that the efficacy of caspofungin against C. albicans was reduced in vivo due to either elevation of chitin levels in the cell wall or acquisition of FKS1 point mutations.

INTRODUCTION

Echinocandins have made a major contribution to antifungal therapy and represent the first class of clinically useful antifungal drugs to inhibit synthesis of the fungal cell wall. The echinocandin target is the catalytic subunit of β(1,3)-glucan synthase, Fks1, which produces β(1,3)-glucan. Echinocandins have fungicidal activity against the majority of Candida spp. and are effective against Candida isolates which are resistant to other antifungal drugs, such as azoles (16, 51, 54). In Aspergillus fumigatus, echinocandins can be cidal or static agents. Although the current incidence of echinocandin resistance is low (53, 55, 56), significant numbers of clinical cases of echinocandin therapy failure have been reported (2, 28, 29, 31, 44, 57, 69). To date, all isolates which acquire resistance to the echinocandins have been shown to have point mutations in the FKS1 gene (51), which generally occur within two hot-spot (HS) regions, HS1 and HS2. The HS1 region of Candida albicans Fks1 (CaFks1) lies between amino acid residues 641 and 649 and HS2 lies between residues 1345 and 1365 (3, 50). Point mutations in the FKS1 gene of Candida spp. have been reported in isolates from patients with breakthrough infections during echinocandin therapy (23, 69). The most frequent point mutation resulting in resistance to caspofungin occurs at Ser⁶⁴⁵ in the HS1 region (36, 50). FKS2 encodes an alternative Fks subunit in Saccharomyces cerevisiae and other fungi (19, 50). Point mutations within FKS2 of C. glabrata can also lead to reduced susceptibility to the echinocandins (23). Some intrinsically echinocandin-resistant fungal species, such as Neurospora crassa, Fusarium solani, Fusarium graminearum, Fusarium verticillioides, and Magnaporthe grisea, have the same echinocandin-resistant Fks1 alleles (32). The prevalence of FKS1 HS mutations was assessed in a bank of 133 Candida spp. with variable echinocandin MIC values collected during 2006 and 2007 from worldwide locations (7). In this study, only 2.9% of isolates contained point mutations in FKS1 HS regions, suggesting that during this period the occurrence of these mutations was low. In vitro, an effect termed “paradoxical growth” or the “Eagle effect” has been observed when some fungal isolates respond less well to high echinocandin concentrations than to lower concentrations of drug (11, 66, 73). One C. albicans clinical isolate which demonstrated the paradoxical growth phenomenon had a 9-fold increase in chitin content at echinocandin concentrations where paradoxical growth occurred (67). This indicates that alterations in cell wall composition and architecture may also lead to reduced susceptibility to caspofungin.

The cell wall is a dynamic structure, and inhibition of one cell wall component can lead to a compensatory increase in another (59). In C. albicans, inhibition of β(1,3)-glucan by caspofungin results in a compensatory increase in chitin synthesis (70). In C. albicans, this increase in chitin synthesis is mediated via multiple cell wall integrity pathways, such as the protein kinase C (PKC), Ca²⁺/calcineurin, and high-osmolarity glycerol (HOG) response pathways (47, 61, 70). Treating wild-type C. albicans cells with Ca²⁺ and calcofluor white (CFW), which activates the Ca²⁺/calcineurin and PKC pathways, respectively, leads to a 3-fold increase in chitin content and reduced susceptibility to caspofungin (47, 70). Cell wall mutants with higher basal chitin contents are also less susceptible to caspofungin (58). Therefore, production of excess chitin has been highlighted to be a possible mechanism of reduced susceptibility to echinocandins in vitro. However, the potential of increased chitin synthesis as an echinocandin resistance mechanism has not been determined in vivo. Here, we investigate the effect of elevated chitin synthesis on in vivo susceptibility to caspofungin in a murine model of systemic candidiasis. We demonstrate that high-chitin C. albicans cells are less susceptible to caspofungin in vivo and that treatment of mice with caspofungin elevates the chitin content of C. albicans cells in infected kidneys. We also present evidence for the in vivo accumulation of fks1 point mutations that are not dependent on prior exposure to echinocandins.

MATERIALS AND METHODS

C. albicans strain and growth conditions.

C. albicans clinical isolate SC5314, which is the most common parent for null mutant construction, was used in this study (21). C. albicans was maintained on Sabouraud agar (1% [wt/vol] mycological peptone, 4% [wt/vol] glucose, and 2% [wt/vol] agar) and incubated at 30°C in a static incubator. Experimental inocula were prepared from YPD (1% [wt/vol] yeast extract, 2% [wt/vol] mycological peptone, 2% [wt/vol] glucose) broth cultures incubated at 30°C overnight with shaking at 200 rpm. For cells with increased chitin content, cells were grown in YPD containing 0.2 M calcium chloride (CaCl₂) and 100 μg of CFW ml⁻¹.

Caspofungin MIC testing: CLSI M27-A2 method.

MICs of caspofungin against C. albicans cells were determined according to the guidelines in CLSI (formerly NCCLS) document M27-A2 using modified (or supplemented) RPMI 1640 broth (48) as described previously (70). Caspofungin concentrations ranged from 0.016 μg ml⁻¹ to 16 μg ml⁻¹.

Cell wall composition analysis.

Cell walls were extracted as described previously (46). Briefly, C. albicans cells were grown in YPD broth with and without Ca²⁺ and CFW for 16 h at 30°C with shaking at 200 rpm. Cells were collected by centrifugation at 3,000 × g for 5 min, washed once with chilled deionized water, resuspended in deionized water, and physically fractured with glass beads in a FastPrep machine (Qbiogene). The disrupted cells were collected and centrifuged at 5,000 × g for 5 min. The pellet, containing the cell debris and walls, was washed five times with 1 M NaCl, resuspended in buffer (500 mM Tris-HCl buffer, pH 7.5, 2% [wt/vol] SDS, 0.3 M β-mercaptoethanol, and 1 mM EDTA), boiled at 100°C for 10 min, and freeze-dried. For quantification of glucan, mannan, and chitin, cell walls were acid hydrolyzed with 2 M trifluoroacetic acid at 100°C for 3 h. The acid was evaporated at 60 to 65°C, and the samples were washed with deionized water and resuspended again in deionized water. The hydrolyzed samples were analyzed by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD) in a carbohydrate analyzer system from Dionex (Surrey, United Kingdom) as described previously (58). The total concentration of each cell wall component was expressed as μg per mg of dried cell wall, determined by calibration from the standard curves of glucosamine, glucose, and mannose monomers, and converted to a percentage of the total cell wall.

Animal models.

Female BALB/c mice (weight range, 18 to 23 g; Harlan Laboratories, Bicester, United Kingdom) were maintained in groups of up to 9 animals per cage and supplied with food and water ad libitum under conditions approved by the United Kingdom Home Office laboratory animals inspectorate. Cells were prepared for injection by growth in YPD or YPD containing Ca²⁺ and CFW under the conditions described above, collected by centrifugation, washed twice, and resuspended in sterile saline (Baxter Healthcare Ltd., Norfolk, United Kingdom).

For the 16-day virulence model, C. albicans (3 × 10⁴ CFU per g body weight) was administrated intravenously (i.v.) via the lateral tail vein of groups of 6 mice. Animals were treated i.v. with daily doses of sterile saline (placebo) or caspofungin solution (1 mg kg⁻¹; Merck Research Laboratories, NJ) for up to 7 days beginning at 24 h postinfection. Caspofungin was diluted in saline and administered as a 200-μg ml⁻¹ solution. Animals were monitored daily for up to 16 days postchallenge. Animals that became immobile or otherwise showed signs of severe illness were humanely terminated by cervical dislocation and recorded as dying on the following day. The kidney burdens and percent change of body weight for each animal were measured as described previously (40–42).

For investigation of early infection stages, a 4-day virulence model was adopted (40, 42); 2 × 10⁴ CFU per g body weight was injected i.v. into the tail veins of groups of 12 mice. The animals were treated i.v. with sterile saline placebo or caspofungin (Merck Research Laboratories) at a daily dose of 1 mg kg⁻¹ for up to 4 days starting 24 h after the day of challenge. Mice from each group (n = 3) were sacrificed each day for 4 days postinfection to investigate disease development and cell wall chitin levels during initial stages of infection. Disease development in each animal was monitored by kidney burdens and body weight change.

To investigate the frequency of acquisition of the FKS1 hot-spot mutation, 10 different high-chitin inocula were prepared as described earlier, with each inoculum used to intravenously infect a female BALB/c mouse (1.3 × 10⁴ to 2.0 × 10⁴ CFU per g body weight). Mice were sacrificed at 24 h postinfection, and kidney organ burdens were determined, with cells from colonies saved for sequence analysis of the FKS1 gene.

Ethics statement.

All animal experimentation in this study conforms to the University of Aberdeen Ethical Review Committee and current United Kingdom Home Office legislation requirements.

Measurement of chitin content by fluorescence microscopy.

Chitin contents were measured as described previously (70). Briefly, to visualize chitin in the C. albicans cell wall, samples of homogenized infected kidneys were fixed with 10% (vol/vol) neutral buffered formalin (Sigma-Aldrich). Fixed cells were stained with CFW (25 μg ml⁻¹), and fluorescence was preserved with Vectashield mounting medium (Vector Laboratories, Peterborough, United Kingdom). All samples were examined by differential interference contrast (DIC) and fluorescence microscopy (456 nm) with a Zeiss Axioplan 2 microscope. Images were captured by a C4742-95 digital camera (Hamamatsu Photonics, Hamamatsu, Japan) and analyzed with the Openlab (version 4.04) software (Improvision, Coventry, United Kingdom). CFW fluorescence was quantified for individual yeast cells using region-of-interest measurements. Mean fluorescence intensities were then calculated, with the number of individual cells analyzed ranging from 96 to 254, and expressed as arbitrary units.

Semiquantitative histopathological analysis.

For animals harvested at 24 h and 72 h postinfection, kidneys were fixed with 10% (vol/vol) neutral buffered formalin and embedded in paraffin, and 5-μm longitudinal sections were cut. Deparaffinized sections were stained with periodic acid-Schiff (PAS) reagent and counterstained with hematoxylin. C. albicans density (number of PAS-positive pixels mm⁻²) and infiltrate density (number of host infiltrate pixels mm⁻²) were measured as described previously (40) as a semiquantitative indicator of fungal burden.

Caspofungin sensitivity testing on agar plates.

Kidneys were harvested from euthanized mice at 24 h and 48 h postchallenge and homogenized with sterile saline, and serial 10-fold dilutions of homogenate were made to determine viable counts. Diluted homogenate was spotted onto YPD plates containing caspofungin at concentrations ranging from 0.031 μg ml⁻¹ to 0.5 μg ml⁻¹ and incubated at 30°C for 24 h. Colonies were counted for each caspofungin concentration, and results were expressed as a percentage of the number of colonies on control YPD plates. The same kidney homogenates were also regrown in YPD medium and used as inocula for broth microdilution susceptibility testing by CLSI method M27-A2.

DNA sequence analysis of FKS1 hot-spot regions.

FKS1 HS regions were analyzed as described previously (50). Briefly, genomic DNA was extracted according to the protocol of Hoffman and Winston (26) from individual colonies from normal- and high-chitin cells in vitro (n = 22), from whole in vitro-grown cultures (n = 3), and from colonies (n = 1 to 120) recovered from homogenized mouse kidneys. HS regions were amplified with the forward (F) and reverse (R) primers (1 μM) 5′-TTT ATT CAA ATT CTT GCC-3′ (HS1-F), 5′-AAT GCC ATG ATG AGA GGT GG-3′ (HS2-F), 5′-GGA ATG CCA TTG TTA TTT CC-3′ (HS1-R), and 5′-GGT ACA GTT TCT CAT TGG CA-3′ (HS2-R) (50). PCRs were performed using Extensor master mix (Abgene) and involved an initial 5-min denaturation step at 94°C, followed by 30 cycles of 94°C for 30 s, 53°C for 1 min, and 68°C for 1 min, with a 5-min final extension step at 68°C. PCR products were purified according to the protocol of Rosenthal et al. (62).

Methods for DNA sequence determinations have been described previously (5). Briefly, purified PCR fragments were sequenced on both strands using 0.7 μM HS-F and HS-R primers. Sequencing reactions were performed with an ABI Prism BigDye Terminator cycle sequencing ready reaction kit (Applied Biosystems) according to the manufacturer's recommendations. Sequences were analyzed with an ABI Prism DNA 3730xl analyzer (Applied Biosystems) by Gene Service, Oxford, United Kingdom. The sequence of FKS1/GSC1 (orf19.2929) from the Candida Genome Database (CGD; http://www.candidagenome.org) was used as a reference (1).

Statistical analysis.

All data sets were analyzed with PASW (version 18.0) statistical software. Organ burdens and body weight changes were compared statistically by the Mann-Whitney U test. Survival data were analyzed by the Kaplan-Meier log rank test. In addition, the chitin measurement by fluorescence intensity was analyzed by the Mann-Whitney U test, with a level of significance between groups set as a P value of <0.05. The data are presented as mean ± standard error of the mean (SEM).

RESULTS

Treatment of C. albicans with Ca²⁺ and CFW increases chitin content but does not affect glucan or mannan levels.

Growth of C. albicans yeast cells in the presence of Ca²⁺ and CFW results in elevated chitin content (47, 70). However, as this treatment could potentially induce other changes in cell wall composition, we investigated its effect on chitin levels of the other cell wall polysaccharides, glucan and mannan. The cell walls of untreated (normal-chitin) and Ca²⁺/CFW-treated (high-chitin) C. albicans cells were extracted, acid hydrolyzed, and analyzed by high-performance liquid chromatography (HPLC; HPAEC-PAD). No significant changes in relative glucan and mannan levels were found upon treatment (Fig. 1B). HPLC data also indicated that growth of cells with Ca²⁺ and CFW resulted in a nearly 3-fold increase in cell wall chitin content (Fig. 1B). Supporting this, fluorescence images of CFW-stained high-chitin cells showed considerably higher intensity than normal-chitin cells (Fig. 1A), indicating elevated chitin content. Analysis of chitin content by fluorescence intensity confirmed an approximately 4-fold increase in high-chitin cells compared to chitin-normal cells (data not shown), consistent with previous work (47, 70).

Fig 1.

Changes in cell wall components when cells were grown in the presence of Ca²⁺/CFW in vitro. (A) Untreated (i) and Ca²⁺/CFW-pretreated (ii) cells showing intensity of CFW staining of nascent chitin in the cell wall. Bars = 10 μm. (B) Cell walls of untreated (normal-chitin) and pretreated (high-chitin) cells were acid hydrolyzed, and released monosaccharide was detected by HPAEC-PAD using a CarboPac PA10 analytical column. Results are expressed as a percentage of dried cell wall (μg mg⁻¹). *, P < 0.05, mean ± SEM (n = 5).

C. albicans cells with elevated chitin content are less susceptible to caspofungin in a murine model of systemic candidiasis.

Prior to the virulence assays, the growth rate and hyphal elongation rates of C. albicans SC5314, which had been pregrown with Ca²⁺ and CFW, were determined. These were comparable to those of untreated cells (data not shown); thus, treatment of cells with Ca²⁺ and CFW had no marked effect on the growth of yeast cells and hyphae.

Mice were infected with chitin-normal and high-chitin C. albicans cells. Caspofungin treatment was initiated at day 1 postinfection and repeated daily for 7 days. All animals of the chitin-normal-infected group treated with caspofungin survived until the end of the experiment (day 16), indicating that caspofungin treatment was successful (Fig. 2A). However, all saline-treated mice infected with chitin-normal cells died within 13 days (Fig. 2A). Kaplan-Meier log rank tests showed these survival differences to be significant (P < 0.05). In contrast, there was no significant difference in the survival of mice inoculated with high-chitin cells whether treated with caspofungin or saline. In addition, the high-chitin-infected mice treated with saline had significantly longer survival than the saline-treated, normal-chitin-infected group (P < 0.05). Kidney fungal burdens in the normal-chitin-infected group treated with caspofungin were reduced by 3 orders of magnitude compared to the saline-treated, normal-chitin-infected group (P < 0.05) (Fig. 2B). In contrast, caspofungin treatment did not reduce kidney fungal burdens of mice challenged with high-chitin cells, and these mice had longer mean survival times (Fig. 2B). Kidney fungal burdens in the high-chitin-infected groups treated with caspofungin were comparable to those of the placebo-treated, normal-chitin-infected group. This suggested that the high-chitin cells were tolerated better than the chitin-normal cells or were less capable of causing disease. There was one example where no fungal cells were detected at the time of death in the kidney of a caspofungin-treated mouse infected with chitin-normal cells. Considered together, the results showed that mice infected with high-chitin cells were able to survive longer than mice inoculated with chitin-normal cells regardless of caspofungin therapy.

Fig 2.

Comparison of the caspofungin response of C. albicans normal- and high-chitin cells in a mouse systemic infection model. (A) Kaplan-Meier survival curves of mice infected with 3 × 10⁴ CFU per g body weight of normal-chitin (NC) or high-chitin (HC) cells. Mice were treated i.v. either with saline or 1 mg per kg of caspofungin (CFN) daily for 7 days (n = 6). (B) Mean fungal burdens of kidneys harvested from each group of mice at the time of death. *, P < 0.05 compared to the normal-chitin plus saline group. Results represent mean ± SEM (n = 6).

According to the survival curves shown in Fig. 2, 50% of mice inoculated with high-chitin cells and treated with caspofungin died by day 3 and the remaining 50% survived until the end of the experiment. To further investigate the caspofungin response of high-chitin cells at the earlier stages of infection, a 4-day systemic infection model was used. Mice were challenged i.v. with normal- or high-chitin cells at 2 × 10⁴ CFU per g body weight, with either caspofungin or saline administered daily from 24 h postinfection. Each day, 3 mice were sacrificed and their kidney fungal burdens and weight change were determined. The kidney fungal burden from mice infected with chitin-normal cells was significantly reduced after the initial dose of caspofungin compared to the saline-treated mice. The kidney fungal burden of chitin-normal-infected mice treated with caspofungin remained significantly lower than that of the other groups for the remainder of the experiment, although no further reduction in burden was measured (P < 0.05) (Fig. 3A). In the other groups of mice, kidney fungal burdens remained essentially unchanged (Fig. 3A). Interestingly, at day 4 postinfection (the black arrow in Fig. 3A), the kidney fungal burden from mice infected with high-chitin cells and treated with caspofungin appeared to increase compared to the other three groups, but this was not statistically significant. The fungal burden results were reflected in altered body weight, where mice infected with normal-chitin cells and treated with caspofungin were the only group not to lose weight over the 4-day virulence model (Fig. 3B). The results suggested that high-chitin cells have a comparable ability to cause disease during early infection, even though in the previous 16-day virulence model the mean survival time of mice infected with high-chitin cells was longer than that for mice infected with normal-chitin cells. In addition, the efficacy of caspofungin against high-chitin cells was compromised, suggesting that the compensatory elevation in chitin was associated with reduced echinocandin susceptibility in vivo.

Fig 3.

Caspofungin responses in vivo. Mice were infected with 2 × 10⁴ CFU per g body weight of normal-chitin (NC) or high-chitin (HC) cells and treated i.v. daily with either placebo or 1 mg per kg caspofungin (CFN), started at day 1 postchallenge. (A) Mice (n = 3) were sacrificed daily, and the fungal load of the kidney burdens was determined. *, P < 0.05 compared to the normal-chitin plus saline group. Results represent mean ± SEM. A black arrow indicates increase of the fungal burden in the high-chitin plus caspofungin group. (B) Weight change of each mouse group. *, P < 0.05, comparing the normal-chitin plus saline group to the others. Results represent mean ± SEM (n = 3).

In vivo caspofungin treatment stimulates chitin biosynthesis in high-chitin cells.

We investigated, by measuring the CFW fluorescence of individual cells, whether the high chitin levels induced during inoculum preparation were maintained during infection and whether caspofungin therapy caused further alterations in chitin levels in vivo (Fig. 4). At day 1 postchallenge, the CFW fluorescence (802 ± 86) of high-chitin cells recovered from infected kidneys was approximately 1.6-fold higher than that of chitin-normal cells (496 ± 31) (P < 0.05) (Fig. 4). Therefore, after 24 h of growth in vivo, high-chitin cells retained an elevated chitin content. Likewise, at 2 days postchallenge, fungal chitin levels were increased in cells recovered from saline-treated, high-chitin-infected animals (963 ± 45) compared to cells from saline-treated, chitin-normal-infected animals (755 ± 215) (Fig. 4). For caspofungin-treated animals, quantitative comparison of chitin levels of cells from infected kidneys could not be performed because too few fungal cells were recovered from kidneys of animals infected with normal-chitin cells. However, the mean fluorescence intensity of fungal cells from high-chitin-infected, caspofungin-treated mice (1425 ± 144) was significantly higher even than that of cells from the high-chitin-infected, saline-treated animals, suggesting that caspofungin treatment stimulated further chitin synthesis in the high-chitin cells in vivo (Fig. 4).

Fig 4.

Chitin contents of cells from infected kidneys. Cells from homogenized kidneys were fixed with 10% formalin and stained with 25 μg ml⁻¹ of CFW. Chitin contents were determined by analyzing CFW fluorescence intensity of individual cells (n = 96 to 254). *, P < 0.05, mean ± SEM (kidneys collected from 3 mice at day 1 and 2 mice at day 2); ‡, no or few cells detected.

Increased colonization of kidney tissue by high-chitin C. albicans cells.

It has been shown that the kidney is the major target organ for C. albicans infection in the mouse i.v. challenge model (43). In addition, the severe kidney damage in mice generally contributes to the overall pathology of disease. Semiquantification of lesions in the kidney has been used to assess differences in the virulence of C. albicans strains (8, 40). We investigated three virulence-associated parameters (the number of lesions, Candida cell density, and immune infiltrate density mm⁻²), assessed by pixel counting, using Photoshop, of color images of stained kidney lesions, to gain a more in-depth understanding of the response of high-chitin cells to caspofungin compared to untreated chitin-normal cells. We used samples collected on day 1 (prior to initiation of therapy) and on day 3 postchallenge (when the greatest difference in fungal burden was observed).

Kidney sections were stained with PAS and hematoxylin to differentiate C. albicans cells and immune infiltrates (Fig. 5). At day 1 postchallenge, both hyphae and yeasts were observed in kidney sections from mice infected with chitin-normal or high-chitin cells (Fig. 5A). In the semiquantitative data, at day 1 postchallenge, there were no statistically significant differences in the number of kidney lesions mm⁻², Candida density, or infiltrate density between mice infected with chitin-normal and mice infected with high-chitin cells (Table 1). The results suggested that kidney infections by chitin-normal and high-chitin cells progress in a similar manner up to 24 h after challenge. At day 3 postinfection, examination of kidney sections from mice treated with caspofungin revealed very few PAS-positive C. albicans elements from chitin-normal-infected kidney lesions compared to those from mice infected with high-chitin cells (Fig. 5B). The mean C. albicans-associated pixel density of high-chitin cells in kidneys of placebo-treated mice increased nearly 5-fold on day 3 compared to day 1 (Table 1). In addition, after 2 days of caspofungin therapy the mean high-chitin cell density was 1.5-fold increased in day 3 kidneys compared to day 1 kidneys (Table 1) and provides further evidence that high-chitin cells were resistant to caspofungin in vivo. This corresponds well with data evaluating kidney fungal burdens determined by the numbers of CFU per g of kidney (Fig. 3A). The mean infiltrate density of kidneys infected with high-chitin cells with and without caspofungin treatment increased 3.1- and 3.5-fold, respectively, from day 1 to day 3 (Table 1). In contrast, at day 3 the mean C. albicans pixel density of the chitin-normal group with caspofungin treatment had decreased 0.6-fold and the mean infiltrate density had increased only 1.4-fold compared to the values at day 1 (Table 1). Without caspofungin treatment, C. albicans and infiltrate densities of the placebo-treated chitin-normal group were increased 28.1- and 8.6-fold, respectively, over the 2-day period (Table 1). Therefore, the efficacy of caspofungin against high-chitin cells was reduced in vivo.

Fig 5.

Histological stained kidney sections of mice infected with normal- and high-chitin cells with or without caspofungin therapy. (A) PAS-hematoxylin-stained kidney sections from day 1 postchallenge with C. albicans and infiltrate densities of each group of mice infected with normal- or high-chitin cells prior to the caspofungin therapy. (B) At day 3 postinfection, stained kidney sections of each group treated with placebo or caspofungin. Bars = 50 μm. Black arrows, PAS-positive fungal cells.

Table 1.

Semiquantitative histological analyses of kidney lesions^a

Day postinfection	Inoculum	Treatment	No. of kidney lesions mm−2	Density (no. of pixels mm−2 [10−4])
				C. albicans	Infiltrate
1	Normal chitin	No	1.89 ± 0.26	0.60 ± 0.09	3.3 ± 0.38
	High chitin		1.71 ± 0.11	0.86 ± 0.13	2.6 ± 0.22
3	Normal chitin	Saline	3.87 ± 0.99	16.78 ± 14.31	28.86 ± 13.38
	High chitin		2.92 ± 0.58	4.09 ± 2.56	9.23 ± 2.64
	Normal chitin	Caspofungin	2.16 ± 0.47	0.34 ± 0.05	4.72 ± 0.99
	High chitin		2.60 ± 0.10	1.30 ± 0.24	8.23 ± 1.39

^aFormalin-fixed and paraffin-embedded kidneys were sectioned and stained with PAS and hematoxylin. Photoshop software was used to analyze PAS-positive fungal and infiltrate materials. Results presented are mean ± SEM, and n is an average of 30 ± 4 regions per kidney.

High-chitin cells recovered from infected kidneys retained reduced susceptibility to caspofungin.

High-chitin cells maintained elevated chitin levels up to 48 h postinfection compared to chitin-normal cells, and chitin content was significantly increased by treatment with caspofungin in vivo (Fig. 4). We next determined the caspofungin susceptibility of C. albicans cells cultured from kidney homogenates by agar plate spot assays. The number of colonies at each caspofungin concentration was counted and used to generate caspofungin dose-response susceptibility curves (Fig. 6). On day 1, prior to echinocandin treatment, fungi from high-chitin-infected mice were less susceptible, with growth unaffected at all caspofungin concentrations tested. In contrast, fungal cells from chitin-normal-infected animals were progressively inhibited by increasing caspofungin concentrations. Fungal cells recovered on day 2 from mice infected with high-chitin cells and treated with caspofungin were not inhibited by any of the caspofungin concentrations tested. Fungi recovered on day 2 from high-chitin-infected, saline-treated mice showed intermediate caspofungin susceptibility (Fig. 6).

Fig 6.

Caspofungin MICs of C. albicans recovered from infected kidneys. Homogenized infected kidney samples were serially diluted and plated on YPD containing a range of concentrations from 0.031 to 0.5 μg ml⁻¹ caspofungin. The number of colonies was counted at each caspofungin concentration, and results are expressed as a percentage of the colonies grown on control YPD plates lacking drug. The results for the normal-chitin cells treated with caspofungin (NC + CFN) group are not shown as insufficient cells were recovered for this test.

Homogenates of kidneys were diluted, regrown in YPD medium, and used as inocula for broth microdilution MIC tests. The overall 50% inhibitory concentrations (IC₅₀s; concentrations resulting in 50% of cell growth compared to the control) of caspofungin for chitin-normal cells recovered from day 1 and 2 were 0.13 to 0.25 μg ml⁻¹ (Table 2). The IC₅₀ for high-chitin cells recovered from saline-treated mice at day 2 was 0.13 μg ml⁻¹. In comparison, the IC₅₀s of caspofungin for high-chitin cells recovered from day 1 and 2 caspofungin-treated animals had increased up to 32-fold compared to the chitin-normal cells (Table 2).

Table 2.

Caspofungin sensitivities of C. albicans cells recovered from infected kidneys^a

Inoculum	Source kidney	CFN IC50 (μg ml−1)
Normal chitin	Day 1	0.13
High chitin		4
Normal chitin	Day 2, saline treated	0.25
High chitin		0.13
Normal chitin	Day 2, CFN treated	—b
High chitin		4

^aCells recovered from kidney burdens were regrown in YPD, washed, and inoculated in RPMI 1640 for caspofungin (CFN) sensitivity testing using the CLSI method.

^b—, insufficient cells were recovered for this test

Caspofungin-resistant cells with elevated chitin acquire mutations in Fks1 hot spot 1.

The FKS1 sequence of cells that had been passaged in vivo was compared with that of cells grown exclusively in the laboratory under chitin-inducing conditions (growth in the presence of Ca²⁺ and CFW). Analysis of the SC5314 inoculum used to infect the mice indicated the presence of heterozygosities in HS2 of FKS1 relative to the reference sequence of FKS1 (CGD). Heterozygous single nucleotide polymorphisms were detected at positions 1410 (Y = C or T) and 1465 (R = A or G) in both the normal- and high-chitin inocula. These were silent mutations. However, at 24 h postinfection high-chitin cells (86%, 12 out of 14) and chitin-normal cells (6%, 1 out of 16) had lost the heterozygosity at these positions.

Treatment of SC5314 with Ca²⁺ and CFW in vitro did not induce alterations in the Fks1 amino acid sequence (Table 3). In addition to sequencing the FKS1 HS regions of individual colonies derived from the in vitro-grown cultures, genomic DNA was isolated from the entire culture; however, no FKS1 HS1 mutations were detected within the entire in vitro population by sequence analysis. This suggests that high-chitin cells in vitro retained reduced susceptibility to caspofungin due to elevated chitin levels in the cell wall, as shown in Fig. 1, rather than alterations in the HS regions of Fks1. However, some FKS1 mutations were found in the in vivo-passaged cultures. Even before initiation of caspofungin therapy, 85% of high-chitin cells (17 out of 20 colonies) recovered at day 1 from infected kidneys were resistant to caspofungin and had an S645Y substitution in Fks1 HS1. Interestingly some mice were infected with a mixed population of fungal cells, with some with an Fks1 S645Y substitution and some without (Table 3). This substitution arose from homozygous FKS1 point mutations, suggesting the possibility of homozygosis of the chromosome. S645Y is a common mutation in Fks1 HS1 associated with decreased sensitivity to echinocandins (50, 51). At day 1, isolates that had acquired the FKS1 point mutation were harvested from 3 different mice infected with high-chitin cells. Isolates with FKS1 point mutations were recovered from 2 mice treated with caspofungin and 2 saline-treated mice at day 2. In contrast, no Fks1 substitutions were found in chitin-normal cells (n = 20 colonies) recovered from infected kidneys. Interestingly, the remaining colonies recovered from infected kidneys that had not acquired the HS1 Fks1 substitution at position Ser⁶⁴⁵ remained sensitive to caspofungin (data not shown). Therefore, there may be variability in the mechanism of echinocandin resistance that evolves when chitin-normal and high-chitin samples are exposed to echinocandins in vivo and in vitro.

Table 3.

Sequence analysis of C. albicans Fks1 hot spots

Day postinfection	Group	Treatment	Mutation
			Hot spot 1 (residues 641-649)	Hot spot 2 (residues 1345-1365)
0	SC5314 (normal chitin)	No	None (0/25)a	None (0/24)
	SC5314 (high chitin)	CaCl2/CFW	None (0/25), (0/10)b	None (0/25)
1c	Normal chitin	No	None (0/20)	None (0/20)
	High chitin	No	S645Y (17/20), (0/120)b	None (0/20)
2d	Normal chitin	Saline	None (0/11)	None (0/11)
		Caspofungin	None (0/1)	None (0/1)
	High chitin	Saline	S645Y (4/6)	None (0/5)
		Caspofungin	S645Y (5/5)	None (0/5)

^aData in parentheses represent the number of colonies that have a mutation in Fks1/total number of colonies tested.

^bData are for samples from 10 inocula (day 0) and 12 individual colonies from 10 separate in vivo experiments (day 1).

^cSamples were collected from three individual mice at 24 h postinfection.

^dSamples were collected from two individual mice at 48 h postinfection.

At day 2 postinfection, all five colonies recovered from the fungal burden plates of mice infected with high-chitin cells and treated with caspofungin had homozygous S645Y HS1 Fks1 substitutions (Table 3). In comparison, 67% (four out of six colonies) of high-chitin cells recovered from the drug-free mice had this mutation. One colony from high-chitin cells recovered from the placebo-treated mouse group with no substitution in HS1 Fks1 had reduced susceptibility and was able to grow in 0.5 μg ml⁻¹ caspofungin (data not shown). Therefore, this HS mutation seems to be selected only in high-chitin cells in vivo in this study.

Further experiments were performed to investigate the frequency of acquisition of FKS1 HS point mutations in high-chitin cells in vivo in 10 parallel cultures that were used to inoculate 10 mice. Conditions of the 4-day infection model were used, except that all mice were sacrificed at 24 h postinfection. Independently grown inocula of high-chitin cells (2 × 10⁴ cells per g body weight) were each injected i.v. into individual mice. At 24 h postinfection, mice were sacrificed and kidney homogenates were cultured in vitro. Genomic DNA was extracted from 12 isolates from the fungal burden plates of each mouse. Genomic DNA was also extracted from the 10 independently grown inocula (∼5 × 10⁹ cells). No point mutations in FKS1 HS1 were found in the in vitro and in vivo samples (Table 3). Therefore, the acquisition of the FKS1 HS1 point mutation was only rarely observed in high-chitin cells in vivo (once in 11 parallel experiments).

DISCUSSION

Echinocandins are the newest class of antifungal drugs introduced into clinical use. They show less toxicity than other systemic antifungal agents and have fungicidal activity against C. albicans (16). In the majority of patients, echinocandin treatment is highly effective; however, in recent years clinical failures have also been reported (44, 57, 69). The majority of these cases involved prolonged exposure to echinocandins and the acquisition of point mutations in the FKS1 gene resulting in echinocandin insensitivity.

Fungi can remodel damaged cell walls by increasing chitin content to maintain cell wall integrity (47, 59, 69). We observed previously that treatment of C. albicans in vitro with sub-MICs of echinocandins could result in a compensatory increase in chitin content (70). This response to echinocandin exposure has also been seen for A. fumigatus (22; L. A. Walker, J. Anjos, K. K. Lee, C. A. Munro, and N. A. Gow, unpublished data). Therefore, elevated chitin biosynthesis is a potential mechanism of resistance or tolerance to echinocandins. As yet there is no direct evidence that this phenomenon leads to clinical failures when treating Candida infections with echinocandins. This study demonstrates that C. albicans cells with increased chitin levels in their cell walls had reduced susceptibility to caspofungin in a mouse model of systemic candidiasis. Therefore, elevated cell wall chitin confers a reduced susceptibility of C. albicans to echinocandins in vivo as well as in vitro.

The cell wall signaling network, including the PKC, HOG, and Ca²⁺/calcineurin pathways, activates CHS expression in response to echinocandins, cations, or CFW (11, 22, 33, 35, 47, 70). In C. albicans, activation of the Ca²⁺/calcineurin and PKC pathways with a combination of Ca²⁺ and CFW leads to a 3-fold increase in chitin content and reduced susceptibility to caspofungin (70). In our experiments, high-chitin cells are primed to activate the cell wall regulatory networks prior to inoculation into mice. Therefore, the basal activity of these pathways may differ between cells with high or normal chitin levels. Because the chitin content of the inoculum was further increased when exposed to caspofungin in vivo, it is clear that these signaling pathways were not fully activated in the cells of the inoculum. A variety of stress pathways have been shown to be activated by echinocandins, including the oxidative stress tolerance pathway, which is important for C. albicans to survive killing by host macrophages (68). It has been demonstrated that Cap1, a key component of the C. albicans oxidative stress adaptation pathway, and Hog1 were activated by 0.19 μg ml⁻¹ of caspofungin (33). In this study, pretreatment of cells with 0.5 mM hydrogen peroxide resulted in decreased sensitivity to caspofungin (33). The heat shock protein 90 (Hsp90) is a key element of the heat stress and other stress responses and plays a critical role in azole resistance by regulating the stability of its client protein, calcineurin (12–14, 64). A synergistic effect of micafungin with an Hsp90 and with a calcineurin inhibitor was observed in isolates harboring the F641S substitution in Fks1 (64). In addition, Hsp90 regulates C. albicans morphology in a temperature-dependent way via the Ras1-PKA (cyclic AMP-protein kinase A) pathway (63). Therefore, echinocandins influence antifungal drug sensitivity and virulence via multiple direct and indirect effects that can involve modulation of a range of stress responses.

“Paradoxical growth,” or the “Eagle effect,” is a phenomenon by which clinical fungal isolates respond less well in vitro to high drug concentrations than to lower drug concentrations (10, 20, 60, 73). The clinical relevance of the paradoxical growth response of C. albicans to echinocandins has not been demonstrated in vivo. A survey of bloodstream Candida isolates, which included C. albicans, C. parapsilosis, C. tropicalis, C. krusei, and C. glabrata, showed that the occurrence of the paradoxical effect is echinocandin specific and species related (9, 27). Furthermore, it has been reported that A. fumigatus infections in vivo can exhibit paradoxical growth in response to echinocandin treatment, with the fungal burdens of caspofungin-treated animals being comparable to those of animals receiving no antifungal treatment. However, this did not translate into a difference in mortality (52). Stevens et al. showed that the in vitro paradoxical effect of a C. albicans clinical isolate was associated with an increase in cell wall chitin content (67). These data support the hypothesis that the echinocandin paradoxical effect may be determined by cell wall chitin content, although the clinical significance of this effect remains to be determined.

The possible role of C. albicans chitin in virulence and host interactions has been addressed previously in a number of studies (reviewed in reference 39). For example, the virulence of the Cachs3Δ chitin synthase mutant with a 70% reduction in chitin content was highly attenuated in a mouse systemic infection model, but the mutant was still able to colonize kidneys (6). Furthermore, isolates with fks1 point mutations have thicker walls, elevated chitin levels, and reduced virulence in a murine candidiasis model and in Toll-deficient Drosophila melanogaster (4). Chitin and β(1,3)-glucan are located in the inner layer of the C. albicans cell wall buried beneath the outer mannan-rich layer (30, 34). However, during in vivo infection and in response to caspofungin treatment, β(1,3)-glucan becomes unmasked in both hyphae and yeast cells, making it more accessible to the β-glucan host receptor dectin-1 (71, 72). Unmasking of β(1,3)-glucan in C. albicans stimulates a strong immune response and increases levels of cytokines such as tumor necrosis factor alpha, interleukin-6 (IL-6), IL-10, and gamma interferon (25, 71). Recent studies have drawn attention to chitin in fungal cell recognition and immune responses (15, 37–39). Chitin and chitin derivatives can act as pathogen-associated molecular patterns and stimulate innate immune cells (37, 38). However, chitin has also recently been shown to block immune recognition of C. albicans by mononuclear cells (45). The immunological effects of chitin seem to also depend critically on the size of the chitin-containing particle (37). The unmasking of β(1,3)-glucan may also lead to increased exposure of chitin on the cell surface and thus contribute to the observed elevation in CFW fluorescence intensity of caspofungin-treated high-chitin cells recovered from infected kidneys (45) (Fig. 4). It is not known whether the high-chitin cells used in this study affect β(1,3)-glucan or chitin exposure or alter immune cell recognition in a way which may result in their attenuated virulence and persistence in the kidneys. However, the finding that chitin may attenuate immune recognition (45) is compatible with the observations reported here. In addition, fks1 mutants with increased chitin contents had a reduced inflammatory response mediated via dectin-1 (4), strengthening the idea that chitin may dampen the innate immune response. This merits further investigation.

Unexpectedly, high-chitin cells but not cells with normal chitin levels acquired FKS1 mutations with a low frequency in vivo in several of the mice tested. Our results are consistent with clinical data that indicate that the acquisition of FKS1 mutations in clinical isolates of Candida species is a rare and stochastic event (7). At 24 h postinfection, there was evidence for selection of homozygous point mutations in FKS1 in high-chitin cells even without exposure to caspofungin (Table 3). Fungal cells recovered at 24 h postinfection from mice infected with high-chitin inocula were also resistant to caspofungin (Table 2). Therefore, in vivo, specific point mutations in FKS1, when they occurred, were rapidly selected and were associated with subsequent reduced echinocandin susceptibility.

A small number of C. albicans clinical isolates with FKS1 point mutations have been tested in the murine models of systemic candidiasis and showed differences in virulence (4, 74). However, clinical isolates are known to vary widely in their ability to cause disease in the systemic candidiasis model (41). Strains selected for in vitro echinocandin resistance that contain an Fks1 Ser⁶⁴⁵ mutation had very few virulence differences, as measured by fungal burdens (65). Therefore, more in-depth studies with higher numbers of isolates and with strains engineered to contain only FKS1 mutations, to rule out other possible mechanisms, are required to determine whether these mutations convey altered fitness in vivo.

Previous studies showed that the K_m of Fks1 from echinocandin-resistant isolates was not altered but the maximum rate of reaction (V_max) was reduced (49). The lower V_max value suggests lower production of β(1,3)-glucan, which may result in activation of the cell wall integrity pathways that induce the compensatory chitin biosynthesis response. Defects in cell wall β(1,3)-glucan due to mutations, such as gas1, are known to activate chitin biosynthesis, and S. cerevisiae fks1 null mutants had increased chitin synthase activity and chitin content (17, 24). Furthermore, C. albicans strains with FKS1 hot-spot mutations have increased chitin content compared to wild-type strains (4, 70). In this study, at day 2 postinfection, high-chitin cells recovered from infected kidneys had increased chitin content in the presence of caspofungin treatment compared to the cells recovered from drug-free mice (Fig. 4). Also, the cell wall composition of high-chitin cells recovered from two individual mice at day 1 postinfection was analyzed by HPLC. These clones harbored an S645Y Fks1 mutation and had higher chitin levels, as measured by HPLC, than normal-chitin cells (unpublished data). An in vitro study showed that an echinocandin-resistant homozygous fks1 point mutant (18) had no further increase in chitin in the presence of caspofungin at a sub-MIC (70). These studies show that there are two nonexclusive mechanisms that can affect echinocandin sensitivity: (i) low-frequency acquisition of FKS1 mutations and (ii) elevation of chitin content via the stimulation of cell wall integrity pathways.

We observed that high-chitin, echinocandin-naive cells that acquired an Fks1 substitution were selected during the first day of infection, sustained a higher chitin content, and were resistant to caspofungin in a systemic candidiasis mouse model. The mean survival time of mice infected with high-chitin cells was considerably longer than that of mice infected with normal-chitin cells. Our experiments also highlight the general conclusion that a detailed understanding of cell wall remodeling processes during infection is critical to decipher host-pathogen interactions at the molecular and cellular level. This in turn may provide valuable insight into improving drug efficacy and identifying drug targets for novel or combination therapies.