|Home | About | Journals | Submit | Contact Us | Français|
Candida albicans is a pathogenic fungus causing vulvovaginal candidiasis (VVC). Azole drugs, such as fluconazole, are the most common treatment for these infections. Recently, azole-resistant vaginal C. albicans isolates have been detected in patients with recurring and refractory vaginal infections. However, the mechanisms of resistance in vaginal C. albicans isolates have not been studied in detail. In oral and systemic resistant isolates, overexpression of the ABC transporters Cdr1p and Cdr2p and the major facilitator transporter Mdr1p is associated with resistance. Sixteen fluconazole-susceptible and 22 fluconazole-resistant vaginal C. albicans isolates were obtained, including six matched sets containing a susceptible and a resistant isolate, from individual patients. Using quantitative real-time reverse transcriptase PCR (qRT-PCR), 16 of 22 resistant isolates showed overexpression of at least one efflux pump gene, while only 1 of 16 susceptible isolates showed such overexpression. To evaluate the pump activity associated with overexpression, an assay that combined data from two separate fluorescent assays using rhodamine 6G and alanine β-naphthylamide was developed. The qRT-PCR results and activity assay results were in good agreement. This combination of two fluorescent assays can be used to study efflux pumps as resistance mechanisms in clinical isolates. These results demonstrate that efflux pumps are a significant resistance mechanism in vaginal C. albicans isolates.
Candida albicans is a pathogenic fungus that causes oral, systemic, and vulvovaginal candidiasis (VVC) (1,–3). VVC caused by C. albicans occurs in at least 75% of women once in a lifetime, and the systemic recurrence of vaginitis caused by Candida species is common (2, 4,–7). The most common drugs for the treatment of this infection are azoles, which target the ergosterol biosynthetic pathway.
Ergosterol is an important component of the cell membrane of C. albicans, as it supports fungal membrane fluidity and proper functioning of the membrane. Many antifungal agents target different components of the ergosterol biosynthetic pathway. For example, terbinafine inhibits squalene epoxidase (Erg1p), fenpropimorph inhibits C-14 sterol reductase (Erg24p), and azoles inhibit lanosterol-14α-demethylase (Erg11p). Erg11p is encoded by the ERG11 gene, an essential gene in the ergosterol biosynthesis (8,–10).
On the basis of the findings of previous studies with oral and systemic isolates, long-term azole treatment and long-term exposure select for azole-resistant C. albicans isolates (11,–13). Several factors can contribute to drug resistance in this organism. C. albicans possesses two families of efflux pumps: ATP-binding cassette transporters (ABC-T) and major facilitator transporters (MFS-T). The genes CDR1 and CDR2 encode ABC transporters, which are overexpressed in the majority of resistant clinical isolates obtained from oral and systemic infections (14). These transporters are energy dependent, and deletion of these genes results in hypersusceptibility to azoles (15, 16). The MFS transporter Mdr1p depends on the proton gradient and transports azoles and other compounds across the plasma membrane. Overexpression of MDR1 has also been identified to be an azole resistance mechanism in oral and systemic clinical isolates (14, 16, 17). Apart from the actions of the efflux pumps, C. albicans can gain azole resistance by overexpression and point mutation of the ERG11 target gene (9, 14, 18, 19).
Recent studies in which efflux pump overexpression in clinical isolates has been documented have used quantitative real-time reverse transcriptase PCR (qRT-PCR) (20). In addition, the fluorescent dye rhodamine 6G (R6G) has been used to quantitate the ABC transporter activity of Cdr1p and Cdr2p (21). While azole efflux is restricted to Cdr1p and Cdr2p ABC-T, R6G is a substrate for many ABC-T (21). In general, the R6G assay has been used in experimental studies. One experimental use of R6G has been for the expression of C. albicans efflux pumps in Saccharomyces cerevisiae to study substrate specificity and to aid in drug development (22).
Recently, a bacterial assay for generalized efflux that uses the fluorescence molecule alanine β-naphthylamide (Ala-Nap) (23) has been applied to Aspergillus fumigatus and C. albicans (24, 25). No oral or systematic clinical isolates have been studied with Ala-Nap.
Azole-resistant vaginal C. albicans isolates have only recently been identified (26). While resistance mechanisms have been studied in oral and systemic isolates, no resistance mechanisms in vaginal isolates have been characterized. There are several possible factors that contribute to azole resistance in C. albicans isolates obtained from the vaginal cavity. First, the bioavailability of fluconazole (FLC) is 12-fold lower in the vaginal cavity than in other tissues (4, 27). This results in colonizing vaginal isolates that are exposed to lower concentrations of FLC after drug administration. Second, the vaginal pH is 4.0, whereas the oral and systemic pH is 7.0; pH is important, as MIC values are pH dependent (2, 28). An understanding of pH dependences has required adjustment of the clinical breakpoints for vaginal MICs (4). Third, the altered pH of the vagina has an effect on Mdr1p pump activity, which is dependent on the proton gradient. Finally, vaginal infections have commonly been associated with a hyperimmune response rather than the overgrowth of C. albicans yeast (29).
Characterization of resistance mechanisms in vaginal isolates required a collection of susceptible and resistant vaginal isolates. The characterization of matched susceptible and resistant isolates from a single patient can contribute significantly to the analysis of resistance, and paired isolates were part of the collection evaluated in the present study. This study describes two fluorescent assays which, when used in combination, measure increased efflux activity resulting from the overexpression of CDR1, CDR2, and MDR1. The study also is the first to describe the molecular mechanisms of azole drug resistance in vaginal C. albicans isolates.
Thirty-eight vaginal clinical isolates of C. albicans were obtained from the Vaginitis Clinic of Wayne State University in Detroit, MI. Isolates were collected from patients with informed consent and Wayne State University Institutional Review Board approval. Of these 38 isolates, 12 isolates represented six matched pairs; i.e., 2 isolates reflecting clinical response and failure, respectively, after FLC therapy were obtained from the same individual. The six matched sets are designated S1-S2, S3-S4, S5-S6, S7-S8, S9-S10, and S11-S12. The matched pairs were verified using multilocus sequence typing (MLST) of four variable housekeeping genes, VPS13, ADP13, ACC1, and AATa, as previously described (30). In brief, genomic DNA was isolated from a single colony of each strain as described previously (31). Defined oligonucleotides were used to amplify each of the four genes using PCR of genomic DNA. While isolates from each of five pairs represented the same strains, isolates S9 and S10 were not same strain of C. albicans (Table 1). To verify that the isolates were C. albicans, they were streaked on CHROMagar plates, and the development of green colonies signified that they were C. albicans. For further identification, the strains were grown in YAD medium (1.7 g of yeast nitrogen base without amino acids and without ammonium sulfate, plus 5.0 g of ammonium sulfate per liter) with 5% xylose (32) to distinguish between C. albicans and Candida dubliniensis. C. albicans grows on YAD medium-xylose, while C. dubliniensis does not (33). Finally, the isolates that had reduced growth in the presence of xylose were verified to be C. albicans by PCR amplification of the cell wall protein-coding gene CRR. The size of the PCR product distinguishes between C. albicans and C. dubliniensis. All isolates were confirmed to be C. albicans. The PCR primers used were CRR-f (5′-GTTTTTGCAACTTCTCTTTGTA-3′) and CRR-r (5′-ACAGTTGTATCATGTTCAGT-3′), as described previously (34). The PCR conditions used were 30 s at 98°C, followed by 30 cycles of 10 s at 98°C, 30 s at 60°C, and 5 min at 72°C and a final extension of 10 min at 72°C. The PCR products were verified on 0.8% agarose gels. All oligonucleotides were ordered from Sigma-Aldrich.
Growth curves were performed to determine the growth rates of the isolates in YAD medium. Altered cell growth of these isolates might affect their susceptibilities to the drugs. Single colonies were suspended in YAD medium plus glucose. After 24 h, the cells at an optical density (OD) of 0.1 were used as the inoculum, and the cells were grown at 30°C for 48 h in a BioTek Take 3 plate reader (BioTek Instruments Inc.). The growth rates of all 38 isolates and control strain SC5314 were similar in YAD medium-glucose (data not shown). All ODs were measured at 600 nm.
The 38 clinical isolates were tested for their susceptibility to the azoles FLC, itraconazole (ITC), and clotrimazole (CLT). FLC, ITC, and CLT were obtained from Sigma-Aldrich. The MICs of the isolates were determined using the CLSI-approved broth microdilution protocol, which determines the MIC to be the concentration of drug that inhibits the growth of the organism by 80% (35). A single colony was suspended in YAD medium with 2% glucose and grown overnight at 30°C. After 24 h, cells at an OD of 0.1 were used as the inoculum. Cells were grown in 96-well plates containing a gradient of drug in twofold serial dilutions. The plates were incubated for 48 h at 35°C. YAD medium with 2% glucose was used for cell growth. YAD medium was used instead of standard RPMI medium, as some of the clinical isolates did not grow well in RPMI medium. A positive growth control lacking drug and a negative growth control lacking cells were included. Cell growth in the wells containing drug was measured by the BioTek plate reader and standardized to the growth of the positive growth control. All ODs were measured at 600 nm.
Two types of fluorescent assays were used to determine the efflux activity in an isolate. First, alanine-β-naphthylamide (Ala-Nap) was used to detect the efflux activities of both types of efflux pumps (ABC-T and MFS-T). Second, rhodamine 6G (R6G) was used to specifically detect the efflux activities of ABC-T.
In preparation for the assay, a single colony of each isolate was suspended in 5 ml YAD medium with 2% glucose and incubated at 30°C with shaking at 180 rpm. After 24 h, cells at an OD of 0.1 were inoculated in YAD medium with 2% glucose and grown at 30°C with shaking at 180 rpm to exponential phase (approximately 6 to 7 h). Cells of each isolate at equal ODs were then washed three times with 1× phosphate-buffered saline (PBS; pH 7), starved for 2 h in 1× PBS, and incubated at 30°C.
The Ala-Nap efflux assay was performed using a previously published protocol (24, 25). Briefly, 100 μl of starved cells was added to a blank 96-well plate (a Corning Costar 96-well black plate; Fisher Scientific). One row of the plate contained only cells, while another row contained both cells and 2% glucose. A stock solution of 0.6 M Ala-Nap (Sigma-Aldrich) was dissolved in ethanol and added to both rows to a final concentration of 0.6 mM. The kinetics of fluorescence were measured for 1 h by using an excitation wavelength of 320 nm and an emission wavelength of 460 nm at 37°C in the BioTek Take 3 plate reader. A well-studied, matched set of oral isolates, isolates 2-76 (#1, susceptible) and 12-99 (#17, resistant) (36), were used as controls in the assay. First, the ratios of the slopes of the efflux in the individual strains in the presence and absence of glucose were calculated for all strains and control strain 2-76. Second, these ratios were compared to the ratio for control strain 2-76 and were plotted on a graph. All experiments were performed in biological triplicate. The slopes were calculated using linear regression on the efflux curves. Errors in the slopes were calculated with the help of the LINEST function in Microsoft Excel software.
The R6G fluorescent dye is specific for the Cdr1p and Cdr2p efflux pumps. The R6G assay was performed as previously described (22, 37). R6G was obtained from Sigma-Aldrich. Briefly, 5 μl of a 10 mM stock of R6G (in ethanol) was added to 5 ml of a starved cell suspension to a final R6G concentration of 10 μM. The cells were incubated at 37°C for 30 min. After incubation, the cells were pelleted for 5 min at 3,700 × g and washed once with 1× PBS. The washed cells were resuspended in 1× PBS plus 2% glucose. Samples (200 μl) were taken from the resuspended cells at 0, 5, 10, 15, and 20 min. The samples were pelleted in a microcentrifuge for 30 s at top speed. The fluorescence of the supernatant was measured using an excitation wavelength of 485 nm and an emission wavelength of 525 nm. Strains 2-76 (susceptible) and 12-99 (resistant) were used as controls. The slopes of fluorescence of each strain, generated by linear regression on the efflux curves, were compared to the slope of strain 2-76, and the results were plotted as the ratios of the slopes. All ODs were measured at 600 nm. All experiments were performed in biological triplicate. Errors in slopes were calculated with the help of the LINEST function in Microsoft Excel software.
Efflux assays were performed on the S. cerevisiae strains in which C. albicans genes were expressed (Table 2, AD mutants). Both Ala-Nap and R6G assays were performed in complete synthetic medium (YAD with 0.77 g of CSM complete powder and 2% glucose per liter; CSM complete medium).
The levels of expression of the mRNA for the genes CDR1, CDR2, MDR1, and ERG11 were measured using qRT-PCR. Exponentially growing cultures of all 38 clinical isolates were prepared from single colonies. RNAs were isolated from the cells using a Qiagen RNeasy minipurification kit following the manufacturer's protocol. The concentrations of RNAs were measured using a NanoDrop plate reader in the BioTek Take 3 plate reader. The A260/A280 ratio was assessed to determine the purity of the RNAs. RNAs with a ratio above 2 were used in the qRT-PCR. As an additional check, 1.2% agarose gels were run, and the appearance of two distinct rRNA bands confirmed the RNA quality. One microgram of RNA from each strain was used to prepare cDNA by using a Thermo Scientific cDNA kit following the manufacturer's protocol. qRT-PCR was performed in an Applied Biosystems 7500 real-time PCR system by the manufacturer's protocol using Bio-Rad iTaq SYBR green supermix. The primers used to amplify the specified genes are listed in Table S1 in the supplemental material. These primers were previously used to determine the levels of expression of mRNA for drug resistance genes in C. albicans (38). Wild-type strain SC5314 was also used as a control. After qRT-PCR was completed, a dissociation step was performed for each reaction to demonstrate that the reactions were specific (data not shown). qRT-PCR for each of the isolates was performed in triplicate. The formula 2−ΔΔCT was used to determine the fold change. First, the threshold cycle (CT) values were compared to the CT values of the actin gene, ACT1. Second, the relative values were compared to the values for SC5314. Errors were calculated as previously described (20, 39). The elongation factor gene CEF3 was used as an additional control. The expression of CEF3 did not vary between the isolates (data not shown). Greater than 2-fold overexpression was defined to be significant; less than 0.5-fold reduced expression compared to the level of expression by SC5314 was defined to be significant.
All data and statistical analyses were performed using GraphPad Prism (version 6.0) and Microsoft Excel software. The types of statistics used are mentioned in the figure legends.
Thirty-eight vaginal C. albicans isolates obtained from 32 women seen at the Wayne State University Vaginitis Clinic were studied. These included 12 C. albicans isolates obtained from 12 women with acute Candida vaginitis in whom no suspicion of clinical resistance existed, and all 12 women responded promptly to conventional fluconazole therapy. Fourteen vaginal isolates of C. albicans were obtained from 14 women with clinically refractory vaginitis and were considered isolates with clinical and in vitro resistance. Finally, 12 C. albicans isolates were obtained from six women from whom two successive C. albicans isolates were available and were studied. All 20 women with resistant C. albicans strains provided a history of significant fluconazole exposure.
In the collection of 38 isolates, a pair of matched isolates was collected from six different patients. To determine if the two isolates in the pair were the same strain, MLST of four housekeeping genes was used (Table 1). In Table 1, the numbers in each cell represent the single nucleotide differences between the isolates initially recovered from the patients (isolates indicated at the top) and the isolates recovered later (isolates indicated in the left column). Five of the six matched pairs and the known matched set 2-76 and 12-99 (36) had no sequence differences between the initial and the subsequent isolates, indicating that both isolates were the same strain. In the sixth matched set (isolates S9 and S10), 2 nucleotide differences were observed.
The MICs of all 38 vaginal clinical isolates to the azole drugs FLC, ITC, and CLT were tested (Table 3). Interpretation of the isolates as susceptible or resistant according to the MICs requires clinical breakpoints. In the present study, the cutoff value for in vitro susceptibility to FLC was ≤1.0 μg/ml, as described previously (4). All isolates with MICs of ≥2.0 μg/ml were considered fluconazole resistant. This is in contrast to CLSI standards, which consider a value of 2.0 μg/ml to indicate susceptibility to fluconazole (40). However, the CLSI breakpoint is based upon clinical predictions for oral and systemic isolates and not vaginal isolates, where the vaginal pH is 4.0 and not the oral and systemic pH of 7.0. Azole activity is dramatically affected by a reduction in pH, and clinical experience indicates that a cutoff of 1 μg/ml is more appropriate (2). Of the 38 vaginal isolates of C. albicans studied, MICs of ≥2 μg/ml were seen in 22, and 19/38 had MICs of ≥4 μg/ml. For both ITC and CLT, clinical breakpoints for oral and systemic isolates have been determined to be 1 μg/ml (14). This breakpoint was used for these vaginal isolates. Using these breakpoints, 22 isolates showed resistance to FLC (MIC ≥ 2 μg/ml), 18 isolates were resistant to CLT (MIC ≥ 1 μg/ml), and 3 isolates were resistant to ITC (MIC ≥ 1 μg/ml) (Table 3). The 3 isolates that were resistant to ITC (isolates S36, S37, and S38) were cross-resistant to both FLC and CLT, while 14 isolates were resistant to both FLC and CLT.
Isolates S2, S4, S8, and S12 from the matched pairs showed an increase in FLC resistance (4- to 32-fold) compared to that for the initial isolates (isolates S1, S3, S7, and S11, respectively) (Table 4), indicating that vaginal C. albicans strains can gain resistance upon exposure to FLC. Isolates S6 and S10 were also FLC resistant, but their initial isolates, isolates S5 and S9, respectively, were resistant as well. Increased ITC resistance was seen in isolates S2 and S4 compared to the susceptibilities of their matched partners, isolates S1 and S3, respectively. CLT resistance was observed in isolates S4 and S12, whereas it was not seen in their matched partners, isolates S3 and S11, respectively (Table 4).
To determine the mechanism of azole resistance, qRT-PCR was used to assess the overexpression of genes associated with drug resistance. qRT-PCR for the four resistance genes CDR1, CDR2, MDR1, and ERG11 was performed on all 38 isolates, including the matched isolates (Fig. 1 and and22).
Twelve of 22 fluconazole-resistant isolates overexpressed CDR1, 10 overexpressed CDR2, 11 showed increased MDR1 expression, and 14 overexpressed ERG11 compared to the levels of expression of these genes in strain SC5314. In susceptible isolates, overexpression was observed only for CDR2 in one isolate (isolate S3). Overexpression of efflux pumps was seen in 72% of the FLC-resistant isolates, 77% of the CLT-resistant isolates, and 2 of the 3 ITC-resistant isolates. ERG11 overexpression was observed in 64% of FLC-resistant isolates, 50% of CLT-resistant isolates, and 3 of 3 ITC-resistant isolates. Several susceptible isolates showed decreases in the levels of expression of the CDR1, CDR2, MDR1, and ERG11 genes (Table 3 and Fig. 1).
The gene expression of the initial and the later isolates in the matched sets was compared (Fig. 2). Among the later isolates, overexpression of CDR1 was observed in isolate S4, overexpression of CDR2 was observed in isolates S4, S8, and S12, overexpression of MDR1 was observed in isolate S8, and overexpression of ERG11 was observed in isolates S4 and S12. No gene overexpression was observed for isolate S2. No overexpression of genes was observed in isolate S6 compared to the level of expression in isolate S5.
To measure efflux pump activity rather than gene expression, two efflux pump assays were employed. The R6G assay measures the activity of ABC transporters, and the Ala-Nap assay measures efflux pump activity in general. A combination of the two assays was tested to determine their utility for measurement of efflux pump activity in resistant isolates. In the R6G assay, R6G is taken up by deenergized fungal cells. With the subsequent addition of glucose, the cells efflux R6G, which is measured in the supernatant. The efflux values are expressed as the slope of R6G efflux from the cells into the supernatant.
Ala-Nap is a dye that is cleaved by proteases within the cell to give fluorescent β-naphthylamide. In the Ala-Nap assay, increased fluorescence indicates a reduction in efflux activity of the pump (24). For each strain, Ala-Nap fluorescence over time is measured in the absence and presence of glucose. An Ala-Nap assay value is the result of a comparison of the slope of the test strain with the slope of the negative control in the presence and absence of glucose. A lower slope ratio means more efflux.
Five S. cerevisiae strains expressing C. albicans genes were used to evaluate the ability of the two assays to measure the efflux activity of the membrane transporters. AD-CDR1 is an S. cerevisiae strain expressing C. albicans CDR1 (CaCDR1) from the endogenous efflux pump Pdr5p promoter. Similarly, AD-CDR2 expresses C. albicans CDR2 (CaCDR2), and AD-MDR1 expresses C. albicans MDR1 (CaMDR1). These strains were constructed in a strain ADΔ background, where all the efflux pumps of S. cerevisiae were deleted (22).
Strain ADΔ showed very little efflux of R6G (see Fig. S1a in the supplemental material) and was used as a negative control for this assay. AD-CDR1 has a high rate of R6G efflux, consistent with CDR1 expression (Fig. 3; see also Fig. S1a in the supplemental material). Similarly, AD-CDR2 has a high rate of R6G efflux, consistent with CDR2 expression. AD-MDR1 has an R6G efflux similar to that of ADΔ, consistent with the fact that MDR1 does not efflux R6G. Strain S288C shows a high level of R6G efflux due to its endogenous expression of PDR5, the S. cerevisiae homologue of C. albicans CDR1.
Strain S288C showed a substantial decrease in the level of accumulation of Ala-Nap, which was 33% of the level of accumulation for ADΔ (100%). Similarly, Ala-Nap accumulation was 38%, 43%, and 77% of that for ADΔ for AD-CDR1, AD-CDR2, and AD-MDR1, respectively. All strains showed significant decreases in their levels of accumulation of Ala-Nap compared to that for the negative control (see Fig. S1b in the supplemental material).
An efflux map (Fig. 3) was generated for the control strains using the results from the Ala-Nap assay (x axis) and the R6G assay (y axis). AD-CDR1, AD-CDR2, and S288C had high R6G values (y axis) and low Ala-Nap values (x axis) and thus graphed in the top right area of the map (Fig. 3). AD-MDR1 had low Ala-Nap values and no increase in R6G values compared to the values for ADΔ. Thus, it graphed to the right of ADΔ, with no increase in R6G (Fig. 3, y axis).
Using the same assays for which the results are demonstrated in Fig. 3 and Fig. S1 in the supplemental material, the efflux activities of the 38 vaginal clinical isolates were analyzed (Fig. 4). Strain 2-76, a susceptible oral isolate from a matched set, was used as a control to which the data were normalized. Strain 12-99, the resistant oral isolate matched to 2-76 that overexpresses CDR1, CDR2, MDR1, and ERG11, was used as a positive control. Strain DSY1050, in which all efflux pumps are deleted, was used as a negative control for pump activity, and SC5314 was included as the wild-type strain.
Most of the resistant strains are located to the right and above compared to the location of strain 2-76 in the efflux map (Fig. 4, red circles and squares). Most of the susceptible isolates (Fig. 4, green circles) did not show an increase in R6G values but varied in their Ala-Nap values.
Susceptible strains S24 and S25 did show high R6G values and low Ala-Nap values (Fig. 4, green squares). This is likely due to overexpression of the CDR4 gene (Table 3), which encodes another ABC transporter. It has previously been shown that CDR4 does not contribute to drug resistance (41). Strains S36 and S37 (Fig. 4, red squares) were located in the top left of the map and had high R6G values. However, the Ala-Nap assay of S36 showed high levels of accumulation (low efflux), while S37 showed no significant change in Ala-Nap levels. This may be due to pumps with altered specificities for Ala-Nap and R6G. In Fig. 4, resistant strains S12, S13, S14, S15, S16, and S38 did not show significant efflux pump activity. Table 3 and Fig. 1 demonstrate that they overexpressed ERG11, which likely contributed to their resistance. In two isolates (isolates S9 and S10), overexpression of MDR1 was observed to correlate with high levels of Ala-Nap activity and low levels of R6G activity (Fig. 4). The resistance mechanism in these two isolates may be due to the overexpression of MDR1.
In the matched sets of vaginal isolates (Fig. 5), the resistant isolate in four out of the five pairs (S2-S1, S4-S3, S6-S5, and S8-S7) mapped above and to the right of its matched susceptible isolate, suggesting that the efflux pumps in these strains contribute to drug resistance. In matched set S8 and S7, the resistant isolate (isolate S8) was to the right but not significantly above S7, suggesting that efflux pumps that do not transport R6G are involved in drug resistance. In the matched set S11 and S12, efflux pumps did not appear to contribute to drug resistance, since S12 was located below its partner, isolate S11, suggesting that the efflux activity of S12 was less efficient than that of its matched partner, isolate S11. Compared to isolate S5, isolate S6 showed no increased FLC resistance (Table 3; Fig. 1) or increased efflux activity (Fig. 5) but did not show increased gene expression, suggesting that S6 expresses efflux pumps unrelated to azole resistance.
The goal of the experiments described in this research was to determine the molecular mechanisms that contribute to drug resistance in vaginal clinical isolates. Thirty-eight vaginal C. albicans clinical isolates collected from the Wayne State University Vaginitis Clinic were confirmed to be C. albicans and had similar growth rates (data not shown), suggesting that their growth rates did not affect the MIC.
Twenty-five of the 38 isolates were resistant to FLC and/or CLT, while 3 isolates were resistant to ITC. The smaller number of ITC-resistant isolates may have been due to the use of ITC clinical breakpoints that are not well defined for vaginal isolates. Clinical breakpoints in the vagina may differ from those in the oral cavity, as described for FLC (2). Alternately, resistance to ITC may be less common in vaginal isolates, which may have implications for treatment. The azole cross-resistance in the isolates (Table 3) is consistent with the overexpression of CDR1, CDR2, MDR1, and ERG11 (Fig. 1) and increased efflux activity (Fig. 4). All azoles are substrates for the overexpressed efflux pumps. Overexpression of ERG11 provides an additional target enzyme that can contribute to resistance.
Increased levels of mRNA (Fig. 1) for the genes encoding the membrane transporters do not always signify increased levels of protein. Therefore, assays that monitor actual efflux activity are a better measure of efflux as a resistance mechanism. In most of the resistant strains, the overexpression of efflux pump genes (Fig. 1) correlated with increased efflux pump activity (Fig. 4). However, qRT-PCR and efflux activity did not correlate in eight strains (S5, S8, S12, S14, S15, S19, S36, and S37). All eight overexpressed CDR1 or MDR1, yet they showed no significant efflux of Ala-Nap and/or R6G. Overexpression of ERG11 may have contributed to resistance in five of these strains (Fig. 1, strains S12, S14, S15, S36, and S37). The increased expression of efflux pump genes without increased efflux activity may be due to altered substrate specificity, protein instability, or translational defects. The disconnect between gene expression and efflux pump activity emphasizes the need for assays that reflect efflux pump activity.
While the efflux activity assays did not detect resistant isolates with ERG11 overexpression, the efflux assays have many advantages. They are straightforward, quick, and cost-effective. Analysis of their data is less complex, and they directly measure efflux activity, which is the dominant resistance mechanism in these vaginal isolates, as well as the dominant mechanism in oral and systemic isolates (14).
The two efflux assays (Fig. 3 to to5)5) showed strong agreement between high efflux rates and azole drug resistance (Fig. 4, red circles). Susceptible strains had less efflux and, hence, in the efflux map were clustered around isolate 2-76 or wild-type strain SC5314. Thus, these efflux assays should be important tools in the armamentarium for the analysis of drug resistance mechanisms.
This study used a collection of matched and unmatched vaginal clinical isolates. Matched isolates have been very useful in the analysis of oral and systemic isolates. Six pairs of matched susceptible and resistant vaginal isolates were obtained in the collection. Using MLST, isolates S9 and S10 (Table 1) were found to not be the same strain, but they had similar MICs indicating resistance. This finding suggests that the patient had previously been exposed to azoles and that the patient was coinfected with two strains or was superinfected with the second strain. Resistant isolate S5 and S6 were the same strain and had high MICs, suggesting that the patient had previously being exposed to azoles. In the other four pairs of matched isolates, the isolates were the same strain and the MIC for the later isolate was increased over that for the initial isolate. In three of these four matched pairs (S2-S1, S4-S3, S8-S7), the later isolate showed increased expression of CDR1, CDR2, or MDR1 by qRT-PCR (Fig. 2) and increased efflux activity (Fig. 5). This implies that long-term treatment with azoles may result in drug resistance by increasing the levels of expression and the activities of efflux pumps.
Azole resistance in C. albicans is most commonly due to the upregulation of genes that enhance efflux pump activity or the upregulation of ERG11 (14; this study). This suggests that resistance might be reversed if drug exposure is withheld, reduced, or eliminated. Hypothetically, a second drug class that reduces the efficiency of efflux pump activity, if used in combination with azoles, can prevent resistance in vaginal as well as oral and systemic candidiasis.
In conclusion, a combined efflux assay that measures efflux activity rather than gene expression was developed. This new assay should be useful in the future to study drug resistance in pathogenic fungi. This assay and qRT-PCR were used to show that the overexpression of membrane transporters and ERG11 contributes to the molecular mechanisms of drug resistance in vaginal isolates, similar to the mechanisms previously seen in oral and systemic isolates.
We thank our colleagues in the T. C. White laboratory for critical discussions and comments on the manuscript. We thank Richard Cannon and Dominique Sanglard for giving us the AD mutant strains and strain DSY1050, respectively.
This research was supported by unrestricted research funds from the School of Biological Sciences, University of Missouri at Kansas City, to Theodore C. White.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.01252-16.