|Home | About | Journals | Submit | Contact Us | Français|
Polyene antifungal drugs, including amphotericin B or nystatin, target ergosterol in the fungal plasma membrane and are used to treat systemic, vaginal and oral fungal infections. In the oral cavity, the available nitrogen sources are primarily in the form of proteins, which are poor nitrogen sources. This study evaluates the effect of protein as a nitrogen source on drug susceptibilities.
Candida albicans was grown in protein [bovine serum albumin (BSA) or casein (CSN)] as a sole nitrogen source, in ammonium sulphate (AS) as a nitrogen source, or in both protein and AS.
Cells grown in BSA or CSN were 4- to 16-fold less susceptible to amphotericin B and nystatin than those grown in AS. Similar results were observed for cycloheximide, but not for fluconazole or caspofungin, and were observed for many C. albicans clinical isolates. The results were observed in two different media, and in broth and on agar. Cells grown under these nitrogen-poor conditions have a reduction in ergosterol sterol levels and a reduction in overall sterol synthesis. Quantitative real-time reverse transcription–polymerase chain reaction analysis shows that some genes involved in sterol biosynthesis are induced under nitrogen-limiting conditions, consistent with the lower sterol levels.
The results demonstrate that nitrogen source has a significant effect on polyene susceptibilities. As these nitrogen-limiting conditions mimic oral nitrogen availability, they suggest that in vitro polyene susceptibilities may overestimate the in vivo susceptibilities to polyene drugs in the mouth.
Candida albicans is a commensal yeast that is part of the normal flora of the mouth, gut and vagina of healthy individuals.1 If normal immune function becomes compromised due to bone marrow or solid organ transplant, or HIV/AIDS, the organism can cause mucosal and systemic disease. As the numbers of immunocompromised patients increase, C. albicans will continue to be a significant cause of morbidity and mortality.2–4
Significant strides have been achieved in elucidating the mechanisms by which C. albicans causes disease, including the identification of adhesion molecules, phospholipases and hydrolytic enzymes called secreted aspartyl proteinases (Saps). Saps are encoded by a 10-member gene family (SAP1–10) and are expressed by the organism in many anatomic sites, including kidneys, oral and vaginal mucosa, CNS, lungs, heart, liver and pancreas.5–8 Many studies have demonstrated the importance of Saps in the pathogenesis of candidiasis in degrading host proteins allowing C. albicans to adhere, invade and destroy host immune molecules.5,6 Recent work has focused on the mechanism by which C. albicans SAP expression is involved in nitrogen acquisition.9,10 In vivo, Candida nitrogen sources are most likely amino acids and proteins from the host. Research from multiple laboratories has shown that SAP expression is repressed in the presence of abundant amino acids or preferred nitrogen sources, such as ammonium.10,11 Thus, SAP expression is not normally found in cells grown in vitro. The fact that SAP expression is detectable in vivo and in reconstituted epithelial models suggests that SAP expression is not repressed by abundant amino acids6,12,13 and that Sap-degraded proteins are a likely source of nitrogen under in vivo conditions.
In vitro drug susceptibility tests are usually performed under artificial growth conditions that supply the organism with easily metabolized carbon (glucose) and nitrogen (predominately glutamine and arginine), conditions which are known to inhibit the expression of virulence-associated genes in vitro.11 Although the effects of antifungal drugs on the expression of virulence factors have been characterized,14 little is known regarding the effect of virulence factors on antifungal drug susceptibility. Culture media that more closely resemble host physiological conditions and allow the expression of virulence factors may have an effect on antifungal susceptibility test results.
In this study, drug susceptibilities were assessed under nitrogen-limiting conditions such as growth in bovine serum albumin (BSA) or casein (CSN). Cells grown under these conditions are characterized by alterations in sterol levels, sterol biosynthesis and sterol biosynthesis gene expression, which might explain these shifts in polyene susceptibilities.
Most studies were performed with the C. albicans reference strain SC531415 and with #1 and #17, a matched set of clinical isolates, azole-susceptible evolving into azole-resistant.16 Clinical isolates 95-120, 95-143, 95-133, 95-175, 97-87, 98-126, 97-197 and 95-165 have been described previously.17 Cells were stored at −80°C in 10% glycerol and were maintained at 30°C.
Several different media were used in this study. Basic cell culture was performed with YEPD (10 g/L yeast extract, 20 g/L peptone and 20 g/L dextrose). Two media were used for nitrogen studies: YNB containing 0.17% yeast nitrogen base without amino acids or ammonium sulphate (AS) and 2.0% dextrose, and YCB containing 1.17% yeast carbon base without AS and 0.1% yeast extract. Neither YNB nor YCB contains a nitrogen source. Therefore, they were supplemented with either 0.2% BSA (+BSA), 0.2% casein sodium salt (+CSN) or 0.5% AS (+AS). In order to determine whether a preferred nitrogen source negated effects of growth in BSA or CSN, AS was added to BSA or CSN (+BSA+AS and +CSN+AS). For all solid media, agar plates contained 15 g/L BactoAgar (BD-Difco).
CLSI MIC assays were performed as follows: single colonies were inoculated into liquid YNB+BSA or YNB+CSN (YNB+BSA/CSN) or YNB+BSA/CSN+AS at 30°C, 180 rpm overnight. Cells were then diluted according to CLSI standards into the media as mentioned earlier, and drug susceptibility tests were performed as per CLSI instructions, with the exception of media and shaking. Cells were incubated and shaken at 35°C, 180 rpm for 48 h. MICs were determined by reading 96-well plates at OD540 and defined as the concentration of drug inhibiting 80% of the growth relative to non-drug treated controls.
Caspofungin and amphotericin B Etests (AB-Biodisk, NJ, USA) were performed as per the manufacturer's instructions with the exception of the media. Single colonies were inoculated into YEPD liquid media and grown overnight at 30°C, 180 rpm. Cells were then diluted to OD600 0.1 and spread, consistent with Etest protocols on YCB+BSA, YCB+BSA+AS, YNB+BSA and YNB+BSA+AS agar media. Etests were incubated at 30°C for 48 h when determinations of MICs were taken.
Drug susceptibility experiments were usually repeated three times, with duplicate samples. Variation between duplicate samples within an experiment was negligible. Experiment-to-experiment variation was minimal, with occasional MICs varying by a single dilution.
The extraction of ergosterol was performed as described previously.18 Individual colonies were picked from a YEPD plate, inoculated into appropriate media and grown overnight at 30°C, 180 rpm. After overnight growth, cells were inoculated at an OD600 of 0.2 and grown for 48 h in relevant media. Cells were pelleted, washed and extracted by potassium hydroxide/ethanol. Cells were not normalized to the optical density of the cell culture. A total of 200 OD units (OD = OD/mL×mL) were used. Data are representative of three independent experiments.
Sterol synthesis was measured, as described previously.19 Cells were grown in media described in the text supplemented with [14C] acetate (sodium salt; 0.1 µCi/µL; 54 mCi/mmol in ethanol; Amersham) and extracted with potassium hydroxide/ethanol and analysed by thin-layer chromatography. Synthesis assay results were confirmed with an additional experiment.
Quantitative real-time reverse transcription–polymerase chain reaction (QRT–PCR) was performed using standard protocols for the following genes: ACT1, ARE2, CAP1, CDR1, CDR2, CDR3, CDR4, CEF3, CRZ1, ERG1, ERG2, ERG3, ERG5, ERG6, ERG9, ERG10, ERG11, FCR1, FLU1, HMG1, HST6, MDR1, NCP1, NDT80, PDR16, RFG1, SSN3, SSN8, TAC1 and UPC2. In brief, cultures were grown in YNB+BSA and YNB+BSA+AS for 6 h. RNA was prepared using the Qiagen RNeasy mini kit (Qiagen, Valencia, CA, USA). Genomic DNA was removed by treatment with RQ1 DNAse followed by heat inactivation. The RNAs were then reverse-transcribed using the Quantitect Reverse Transcription Kit (Qiagen). The cDNA was subsequently analysed in quantitative PCR using a SYBR green mastermix in an ABI 7500 (Applied Biosystems, Foster City, CA, USA), following the manufacturer's recommended protocols. Oligonucleotides for the 28 genes associated with resistance and for the 2 control genes (ACT1 and CEF3) used for the QRT–PCR have been described previously.20 The oligonucleotides for the SAP genes are as follows: SAP2F: 5′ TCCTGATGTTAATGTTGATTGTCAAG 3′; SAP2R: 5′ TGGATCATATGTCCCCTTTTGTT 3′; SAP4F: 5′ CAATTTAACTGCAACAGGTCCTCTT 3′; SAP4R: 5′ AGATATTGAGCCCACAGAAATTCC 3′; SAP5F: 5′ CATTGTGCAAAGTAACTGCAACAG 3′; SAP5R: 5′ CAGAATTTCCCGTCGATGAGA 3′; SAP6F: 5′ CCTTTATGAGCACTAGTAGACCAAACG 3′; SAP6R: 5′ TTACGCAAAAGGTAACTTGTATCAAGA 3′. Data are representative of RT reactions derived from three, independently grown cultures.
In order to investigate the effect of nitrogen acquisition on antifungal drug susceptibility, assays were performed under previously described Sap-inducing conditions.21 Cells were grown in YNB+BSA, YNB+AS and YNB+BSA+AS (Figure 1). Cells grown in media with BSA as the sole nitrogen source exhibited an MIC80 to amphotericin B of 1.1 mg/L. Cells grown in YNB+AS or YNB+BSA+AS had MIC80 8- to 4-fold lower than BSA alone.
In an effort to determine whether this phenomenon was directly related to the amount of an easily metabolized nitrogen source, increasing amounts of AS were added to YNB+BSA. As the amount of AS increased, the MIC of amphotericin B decreased (Figure (Figure22).
In order to ensure that the reduced susceptibility to amphotericin B in YNB+BSA media was not due to a pleiotropic effect specific to BSA, CSN was used as a protein nitrogen source (Table 1). Growth in either YNB+BSA or YNB+CSN resulted in an 8-fold increased MIC to amphotericin B relative to YNB+BSA+AS or YNB+CSN+AS.
To determine whether the phenomenon was specific to amphotericin B or was a more general characteristic of polyenes, another polyene, nystatin, was tested. Growth in YNB+BSA or YNB+CSN reduced the susceptibility to nystatin, relative to growth in YNB+BSA+AS or YNB+CSN+AS. However, the reduced susceptibility to nystatin was less than the reduction for amphotericin B. For nystatin, a 4-fold increase was detected in YNB+BSA and 2-fold in YNB+CSN.
A small but reproducible 2-fold reduction in susceptibility was also detected to fluconazole in cells grown in BSA alone. However, no change in susceptibility was detected in CSN media relative to CSN media supplemented with AS.
Previous studies demonstrated that strains with mutations in the ergosterol biosynthetic pathway exhibited altered polyene susceptibility and an increased susceptibility to cycloheximide.22 Cells grown in YNB+BSA and YNB+BSA+AS were tested for altered susceptibility to cycloheximide. Cells grown in YNB+BSA exhibited a 64-fold increase in cycloheximide susceptibility when compared with cells grown in YNB+BSA+AS.
It was possible that decreased susceptibility to amphotericin B in YNB+BSA media was specific to YNB-based media. Previous studies have shown that SAPs are expressed in YCB media as well as YNB media.23 To demonstrate that the change in amphotericin B MIC was associated with all Sap-inducing conditions and not specific to YNB media, YCB media was also tested. Strain SC5314 exhibited an 8-fold increase in amphotericin B resistance when grown in YCB+BSA relative to cells grown in either YCB+AS or YCB+BSA+AS (Table (Table2).2). To ensure that the change observed in MIC to amphotericin B was not specific to strain SC5314, clinical isolates #1 and #17 were also tested in YCB media. As with SC5314, the two clinical isolates exhibited 8-fold increases in amphotericin B MICs when grown in YCB+BSA when compared with YCB+AS or YCB+BSA+AS. As the increase was observed for #17 as well as #1, it is clear that azole resistance mechanisms, which are found in #17, are not associated with this change in polyene susceptibility.
Caspofungin is a new drug in the antifungal armamentarium that targets glucan synthase and for which Etests are available. To determine whether growth in protein media resulted in reduced caspofungin susceptibility, cells were grown on two different agar media with protein as a nitrogen source. Limited differences, 1.5- to 3-fold, were observed in the caspofungin susceptibility of three C. albicans isolates grown on media with YNB+BSA or YCB+BSA relative to cells grown on YNB+BSA+AS or YCB+BSA+AS (Table (Table3).3). As caspofungin contains several peptide bonds, it was of interest to test the effect of Sap on caspofungin. Experiments in which Sap-containing culture supernatants were incubated with the drug caspofungin or with cells did not alter the MIC (data not shown). This suggests that Sap proteinase activity does not degrade caspofungin.
In order to determine whether alterations in amphotericin B susceptibility were limited to the laboratory strain SC5314 and the matched clinical isolates #1 and #17, additional clinical isolates were used in amphotericin B susceptibility tests in the presence and absence of AS. Strains were chosen that were isolated from different anatomic locations and varied in fluconazole susceptibility, ranging from susceptible to resistant.17 Isolates 95-120 and 98-143 are fluconazole-resistant, systemic isolates. Isolates 97-87, 98-126 and 97-197 are azole-susceptible, systemic isolates. Strains 95-133 and 95-175 are resistant and susceptible oral isolates, respectively. Strain 95-165 is a resistant vaginal isolate. In all cases for all the strains tested, cells grown in the absence of AS were 5.3–15.8-fold less susceptible to amphotericin B relative to cells grown in the presence of AS (Table (Table4).4). A collection of amphotericin B-resistant isolates is not available.
To determine whether the increase in MIC to polyene drugs coincided with a drop in total sterols, levels of ergosterol were quantified in cells grown in YNB+BSA and YNB+BSA+AS (Figure (Figure3).3). Ergosterol was extracted from cell cultures at 24 and 48 h, as previously described.18 Cells grown in YNB+BSA at 24 and 48 h contained less ergosterol relative to cells grown in YNB+BSA+AS. At 24 h, the ergosterol content of cells grown in YNB+BSA was 63.0% of the ergosterol content of cells grown for the same time in YNB+BSA+AS. At 48 h, the ergosterol content of cells grown in YNB+BSA was 64.0% of the ergosterol content of cells grown for the same time in YNB+BSA+AS. These data demonstrate that the increase in MIC to polyene drugs correlates with a reduction in cellular ergosterol.
Data from C. albicans and Candida lusitaniae suggest that cells with defects in the ergosterol biosynthetic pathway prevent the synthesis of ergosterol, which is the target for polyene drugs.22,24 To determine whether the change in MIC to polyenes and the reduction in cellular ergosterol are the result of decreased synthesis and/or a reduction of stored ergosterol, cells were grown in the presence of 14C-labelled acetate in YNB+BSA, YNB+CSN, YNB+BSA+AS or YNB+CSN+AS for a 3 h pulse starting at 45 h and harvested at 48 h of growth (when MICs are normally determined). Cells in YNB+BSA exhibit total sterol synthesis rates that are 20% of the AS controls and ergosterol synthesis rates that are 18% of AS controls. Similar results were obtained with CSN media, in which both total sterol and ergosterol levels are reduced to 18% of the AS supplemented controls (Figure (Figure44).
QRT–PCR was used to monitor the expression of a collection of 28 genes associated with resistance as well as four SAP genes and two control genes (ACT1 and CEF3). QRT–PCR was performed on RNA from cells grown in YNB+BSA and in YNB+BSA+AS in three biological replicates. The results were normalized to ACT1, with a focus on genes that are overexpressed or underexpressed at least 2-fold. The results were similar when the expression patterns were normalized to CEF3, instead of ACT1.
Of the 32 genes, SAP2 increased 87-fold in YNB+BSA media compared with YNB+BSA+AS media (P < 0.5). A limited number of genes associated with resistance were reproducibly overexpressed but were not statistically significant, including two major facilitator pumps MDR1 (29-fold) and FLU1 (4-fold) and three ergosterol genes ERG11 (8-fold), ERG2 (4-fold) and ERG6 (3-fold). The reproducible overexpression of the pumps and of the sterol genes is consistent with the changes in MIC, sterol levels and sterol synthesis rates described earlier, suggesting that the phenotype is associated with increased drug efflux by major facilitator genes, and/or by changes in the later part of the ergosterol pathway. No change (1.07-fold) was observed for UPC2, the transcription factor that regulates genes in the ergosterol pathway.25,26
Recent studies have investigated C. albicans' ability to acquire and import protein as a nitrogen source. Despite the knowledge that host proteins are a likely nitrogen source for Candida, antifungal susceptibility assays are routinely performed in the presence of nitrogen sources that are easily acquired and preferentially utilized by the fungus. This study has analysed the effect of protein as a nitrogen source on the susceptibility of C. albicans to several antifungal drugs. Results indicate that C. albicans susceptibility to antifungal drugs, particularly polyenes, is reduced when proteins are the sole nitrogen source for cell growth. Previous studies have demonstrated increased resistance to polyene drugs in strains with defects in the biosynthesis of ergosterol, which results in reduced amounts of ergosterol, the target of polyenes.22,24 In order to determine whether increased resistance to polyenes was a consequence of reduced cellular ergosterol, ergosterol levels and synthesis were determined and found to be lower in cells grown in BSA or CSN media relative to the same media supplemented with AS, a preferred nitrogen source. Additionally, the expression of ERG genes and efflux pumps was increased in cells grown in BSA or CSN media, relative to cells grown in either media supplemented with AS. Increased expression of ERG genes is consistent with decreased levels of ergosterol and total sterols, as demonstrated by Song et al.19 Despite the increase of major facilitator expression, there was only a modest decrease in azole susceptibility (Table (Table11 and data not shown).
AS, an easily metabolized nitrogen source, increased the susceptibility of C. albicans to several antifungal drugs when it was used to supplement media where proteins were the sole nitrogen source. The effect was largely seen for the polyene drugs amphotericin B and nystatin. However, absolute susceptibilities differed between liquid and agar media tests, as well as between YNB and YCB media. Much of this may be a consequence of modifying the protocols for both CLSI and Etest in order to utilize protein as a nitrogen source. Both protocols are optimized for RPMI media, which has abundant amino acids. Some amino acids, such as proline, are poor nitrogen sources, whereas others, such as glutamine and arginine, are easily metabolized. When RPMI, which contains significant glutamine and arginine, was supplemented with proline, AS, BSA or combinations, there was no change in the MIC to polyene drugs (data not shown), suggesting that the amino acids in RPMI serve as rich nitrogen sources. The unchanged polyene MIC in RPMI supplemented with BSA also suggests that BSA does not bind to amphotericin B and inhibit its antifungal activity.
The addition of AS to whole BSA or CSN media serves an additional purpose. It is possible that BSA protects C. albicans from polyene drugs by binding directly to polyenes, preventing the drugs from interacting with ergosterol. It is also possible that BSA may bind to ergosterol, preventing polyenes from binding. The addition of AS to the media, which increases the susceptibility of Candida, supports the conclusion that BSA is not binding to amphotericin B, or preventing it from binding directly to the cell surface. To demonstrate further that BSA was not providing a protective effect against polyenes, CSN was used as the sole nitrogen source. Again, protein as a nitrogen source increased the MIC to polyene drugs. The addition of AS to CSN media abrogated the increase in MIC, consistent with results obtained with BSA. This suggests that nitrogen source, not protein binding to drug, is responsible for altered polyene MICs. This conclusion is supported by caspofungin results. Caspofungin has been reported to be highly bound by serum.27 Despite caspofungin being bound by protein, its MIC is not greatly altered by the presence or absence of AS. These data provide further support that altered polyene susceptibility is the result of a different nitrogen source.
CAS amino acids, an enzymatic degradation of CSN, were also used as a nitrogen source (data not shown). Cells grown in YNB media+CAS amino acids and YNB+CAS amino acids+AS had similar MICs of amphotericin B, suggesting that CAS amino acids also act as an easily metabolized nitrogen source. The change in MIC associated with growth in CSN (whole protein) relative to CAS amino acids (degraded protein) suggests that growing in media where protein is the sole nitrogen source, degrading those proteins and importing the resulting peptides, alters Candida's physiology, resulting in reduced susceptibility to amphotericin B.
The mechanism by which peptides are assimilated and Saps are expressed has recently been investigated.9,10 Cells grown in BSA as the sole nitrogen source sense the presence or absence of extracellular amino acids via Csy1p. In low amino acid conditions, such as growth on BSA, an inactive transcription factor Stp1p is proteolytically cleaved and translocates to the nucleus, binding to the promoters of genes such as SAP2 and oligopeptide transporters, thus increasing their expression. Under conditions where abundant, easily metabolized nitrogen sources are present; another transcription factor, Stp2p, is activated and increases the expression of amino acid permeases. Under the same conditions, the transcription factor GAT1 increases the expression of the general amino acid permease, GPA, in order to import extracellular amino acids.11 This abundance of amino acids represses the expression of Saps, conditions conferring normal susceptibility to amphotericin B.
Despite the connections between GAT1 expression, nitrogen catabolite repression and Sap activity, connecting these events with ergosterol biosynthesis and amphotericin B susceptibility is more difficult. One potential connection is the levels of acetyl CoA in the cell. It is possible that acetyl CoA serves as a central link to many of the processes described earlier. It is an initial substrate for the biosynthesis of amino acids, amino acid derivatives and sterols and can serve as a product in the degradation of peptides and amino acids for nitrogen (and carbon). Cells grown on protein as a nitrogen source might be limiting in the levels of acetyl CoA. Depending on the cell's needs, limiting amounts of acetyl CoA might be directed to amino acid biosynthesis and directed away from other processes such as sterol biosynthesis. Reduced flow through the sterol biosynthetic pathway would result in reduced sterol synthesis, reduced ergosterol and other sterols, increased expression of ERG genes and finally an altered polyene susceptibility.
Standardized in vitro MIC testing is an important clinical tool that is a good measure of how fungal cells respond to drug. In general, MICs correlate well with cell response to drug in the host. However, MIC testing was not designed to mimic the growth conditions in the host. This study demonstrates one facet of the environment in the host that is not reflected in standard MIC testing. The results are presented not as a recommendation to modify standard MIC testing, but to further understand the interactions between the fungal cells, drugs and the host. Only when we fully understand all three components of this interaction can we truly control these infections.
This research was funded by NIH NIDCR R01 grants DE-11367, DE 14161 and DE17078. B. G. O. was supported by NIH pathobiology training grant T32 AI 07509.
None to declare.
We thank the members of the White Laboratory for their comments regarding this manuscript.