Comparison of Human and Soil Candida tropicalis Isolates with Reduced Susceptibility to Fluconazole

Infections caused by treatment-resistant non-albicans Candida species, such as C. tropicalis, has increased, which is an emerging challenge in the management of fungal infections. Genetically related diploid sequence type (DST) strains of C. tropicalis exhibiting reduced susceptibility to fluconazole circulated widely in Taiwan. To identify the potential source of these wildly distributed DST strains, we investigated the possibility of the presence in soil of such C. tropicalis strains by pulsed field gel electrophoresis (PFGE) and DST typing methods. A total of 56 C. tropicalis isolates were recovered from 26 out of 477 soil samples. Among the 18 isolates with reduced susceptibility to fluconazole, 9 belonged to DST149 and 3 belonged to DST140. Both DSTs have been recovered from our previous studies on clinical isolates from the Taiwan Surveillance of Antimicrobial Resistance of Yeasts (TSARY) program. Furthermore, these isolates were more resistant to agricultural azoles. We have found genetically related C. tropicalis exhibiting reduced susceptibility to fluconazole from the human hosts and environmental samples. Therefore, to prevent patients from acquiring C. tropicalis with reduced susceptibility to azoles, prudent use of azoles in both clinical and agricultural settings is advocated.


Introduction
Due to the elevated number of risk populations, the prevalence of fungal infections has increased significantly in past two decades. Candida species are the most frequently isolated fungal pathogens causing morbidity and mortality in seriously immunocompromised hosts. Although Candida albicans is the most prevalent species in hospitalized individuals and in nosocomial infections, there has been a shift toward the more treatment-resistant non-albicans Candida species [1,2,3,4,5]. This has become an emerging issue in the management of fungal infection. The prevalence of these species differed significantly in various geographic areas [4,5,6,7,8,9,10]. Candida glabrata was the most frequently isolated species in Western countries [4,11,12], whereas C. tropicalis predominated in Asia [5,13,14,15]. For the treatment, azoles, echinocandins, polyenes, and 5-flucytosine are the four major classes of antifungal drugs. Due to low cost and less side effects, fluconazole has become one of the most commonly prescribed drugs.
Among the phenomena associated with azole resistance, 'trailing' describes the reduced but persistent growth that some isolates exhibit at drug concentrations above the minimum inhibitory concentrations (MICs) in broth dilution tests [16]. When the MIC of an isolate measured after 48 hours (h) incubation is approximately 4-fold higher than that at the 24 h point [17], the isolate is defined to have trailing growth. Thus, in the present study, isolates with fluconazole MICs$64 mg/L or voriconazole MICs$4 mg/L after 48 h incubation were considered to have reduced susceptibility to azole drugs.
In order to monitor the trends of species distribution and drug susceptibilities of yeast pathogens, the Taiwan Surveillance of Antimicrobial Resistance of Yeasts (TSARY) program was initiated in 1999 [18]. Subsequently, two more rounds of TSARY were conducted in 2002 and 2006. Previously, we found 23 of the 162 C. tropicalis isolates collected from TSARY in 1999 with fluconazole MICs$64 mg/L to be closely related despite being collected from different hospitals throughout Taiwan [19,20]. Furthermore, 5 of the 23 isolates exhibiting reduced susceptibility to fluconazole were from hospital N4 and all belonged to the same diploid sequence type (DST), DST140 (allele combination, 1, 3, 3, 17, 54, and 3 for ICL1, MDR1, SAPT2, SAPT4, XYR1, and ZWF1a, respectively). However, we did not find any evidence to suggest that these five isolates were transmitted horizontally from person to person [21]. In addition, 2 genetically closely related DST C. tropicalis strains, DST140 and DST98 (allele combination, 1, 3, 3, 17, 9, and 3), exhibited reduced susceptibility to fluconazole were identified in TSARY 1999 and 2006. Among them, 18 DST140 isolates were recovered from 10 different hospitals located in all 4 geographic regions in Taiwan and 7 DST98 ones were recovered from 4 different hospitals, two each in northern and southern Taiwan. There were also 3 DST149 (allele combination, 1, 44, 3, 7, 58, and 3) isolates exhibiting reduced susceptibility to fluconazole and from 3 hospitals located in northern Taiwan were identified. These results indicated that those DST strains exhibiting reduced susceptibility to fluconazole circulated widely in Taiwan from 1999 to 2006 and their presence was not a result of outbreaks in certain hospitals or geographic regions [21].
Exposure to azole compounds paves the way for the selection and enrichment of fungal isolates exhibiting reduced susceptibility to drugs, which may occur in patients receiving azole treatments. However, emergence of person to person transmission of isolates exhibiting reduced susceptibility to fluconazole during medical treatment is unlikely. Alternatively, the use of azole compounds in the environment may select organisms exhibiting reduced susceptibility to drugs, which may find their ways to human. Candida tropicalis is prevalent in organically enriched soil and aquatic environments [22] as well as in wild birds [23]. Thus, in the present study, we investigated the presence in soil of these C. tropicalis strains exhibiting reduced susceptibility to fluconazole, especially DST98, DST140, and DST149.

Candida tropicalis isolates
A total of 56 C. tropicalis isolates (

Molecular typing
To determine whether the C. tropicalis isolates with reduced susceptibility to fluconazole recovered from soil samples were genetically related to those from human, we applied pulsed field gel electrophoresis (PFGE) method to analyze their genetic relatedness. The 18 isolates that were investigated included 11 soil and 7 human isolates. The soil isolates comprised 6 isolates exhibiting reduced susceptibility to fluconazole (NHUE10, 28, 33, 40, 48, and 56) and 5 susceptible ones (NHUE11, 27, 36, 42, and 52). The 7 clinical isolates (YM990275, 490, 537, 579, 592, 649, and 659) all exhibited reduced susceptibility to fluconazole. As expected, all 6 DST140 isolates from human shared an indistinguishable PFGE pattern (Fig. 1). In addition, 2 isolates exhibiting reduced susceptibility to fluconazole from soil, NHUE48 and NHUE56, shared the same PFGE pattern with those human isolates. Among the remaining 4 soil isolates exhibiting reduced susceptibility to fluconazole, 3 (NHUE28, NHUE33, NHUE40), from different samples, shared at least 97% similarity. Furthermore, fluconazole susceptible isolates recovered from soil were not closely related to the ones exhibiting reduced susceptibility to fluconazole from the same areas, i.e. NHUE 11 vs. NHUE 10, NHUE 36 vs. NHUE 33, and NHUE 42 vs. NHUE 40. The fact that YM990579, a DST149 human isolate, also shared the indistinguishable PFGE pattern with the six DST140 human isolates, suggests that MLST has a higher discriminatory

Growth curves
All 18 soil isolates with fluconazole MICs$64 mg/L after 48 h incubation, such as NHUE40, NHUE48, YM9900579, and YM990649, were able to grow in the presence of 64 mg/L fluconazole (Fig. 2). In contrast, the growth of the susceptible ones  was inhibited, such as NHUE42 (Fig. 2, triangles solid line) and ATCC13803 (Fig. 2, crosses solid line). Hence, those isolates with fluconazole MICs$64 mg/L definitely had reduced susceptibility to fluconazole than those isolates with lower fluconazole MICs.

Susceptibilities to drugs used in agriculture
Since 17 of the 18 soil isolates exhibiting reduced susceptibility to fluconazole also had high voriconazole MICs, we selected 6 isolates to investigate whether isolates with reduced susceptibility to fluconazole from human and soil also had reduced susceptibility to other azole drugs used in agriculture. The relative growth of 6 isolates after incubation for 48 h is showed in Table 2. In the presence of 64 mg/L fluconazole, isolates exhibiting fluconazole reduced susceptibility, NHEU40, NHUE48, YM990579, and YM990649 grew better (66-100% of growth) than the fluconazole susceptible ones, ATCC13803 and NHUE42 (27-32% of growth), confirming our previous results obtained from broth microdilution method. All tested isolates grew poorly (10-11% of growth) in the presence of 64 mg/L penconazole and tebuconazole. Interestingly, the 4 isolates exhibiting fluconazole reduced susceptibility grew better than fluconazole susceptible ones in the presence of 4 mg/L penconazole (36-60% vs. 23-24% of growth), 4 mg/L tebuconazole (54-83% vs. 12-15% growth), 16 mg/L fluquinconazole (71-87% vs. 41-59% of growth), or 64 mg/L tridimenol (24-35% vs. 15-17% growth). Hence, isolates exhibiting fluconazole reduced susceptibility were also less susceptible than the fluconazole susceptible ones to other azole drugs used in agriculture.

Discussion
Candida tropicalis is one of the most frequently isolated nonalbicans Candida species from patients [8,10,24,25]. Furthermore, C. tropicalis develops drug resistance in the presence of fluconazole much more rapidly than other Candida species [26,27]. Our findings in the present study suggest that soil may be a potential source of C. tropicalis with reduced susceptibility to azole type of antifungal drugs.
The observation that 3 C. tropicalis isolates (NHUE28-30) with reduced susceptibility to fluconazole recovered from the same soil sample, #24, belonged to the same DST strain (DST149) suggests that they may be the progenies of the same strain. Nevertheless, one isolate (NHUE31) recovered from the same #24 soil sample was susceptible to fluconazole. On the other hand, of the 7 isolates (NHUE37-43) recovered from another soil sample, #6, 5 exhibited reduced susceptibility to fluconazole and belonged to 4 different DSTs (DST183, 187, 149, and 227). These results indicated that multiple strains of C. tropicalis exist in a single soil sample.
Long duration of drug exposure and high numbers of reproducing microorganisms contribute to the selection and/or enrichment of drug-resistant individuals. In fungi, drug resistance is more likely to be the outcome of sequential accumulation of adaptive mutations in the chromosomes [28,29]. Azole-resistant isolates have been shown to emerge following microbial exposure to the drugs in patients and in agricultural settings. The observation that azole-resistant Aspergillus fumigatus isolates recovered from soil and compost were genetically related to clinical resistant isolates suggests the existence of an environmental route for developing drug resistance in fungi [30,31,32]. Similar to some antibiotic resistance in bacteria [33,34,35], fluconazole resistance in fungi also involves step-wise mutations affecting more than one  genes [36,37,38]. Therefore, the C. tropicalis isolates from soil exhibiting trailing growth phenomena in the present study are prone to be resistant.
In the present study, we have found that C. tropicalis isolates with fluconazole MICs$64 mg/L recovered from human and from soil, albeit from different geographic regions of Taiwan, shared indistinguishable PFGE patterns and belonged to the same DSTs. Whether the original development of C. tropicalis exhibiting reduced susceptibility to fluconazole occurred in human hosts or in the environments needs further investigation. However, our observation is consistent with the idea that medically important drug resistant isolates, not only bacteria but also fungi, exist in the environments.
Previous study has shown that in human immunodeficiency virus infected patients, C. albicans and other yeast species were cross-resistant to medical and agricultural azole drugs [39]. However, no C. tropicalis was recovered in that study. The observation that isolates with reduced fluconazole susceptibility were also more resistant than the susceptible ones to fluquinconazole and tridimenol but not to penconazole and tebuconazole ( Table 2)  Since both efficient reproduction and spreading of resistant fungi in the environment are to be anticipated, we are confronted with the major challenge of drug-resistance development in environmental setting on a global scale. Hence, to reduce and prevent patients from acquiring azole-resistant C. tropicalis and other fungal species, it is advisable to take precaution against unnecessary use of azoles in not only clinical but also agricultural settings. The mechanisms contributing to the reduced susceptibility to azole drugs are under investigation. Further study should be performed to compare C. tropicalis isolates recovered from soils in the fields as well as those from farmers for genetic relatedness and susceptibility to various antifungal drugs.

Strain isolation
To isolate yeasts from the soil, we transferred approximately one gram of soil from each sample into a tube containing nine ml of sterile water and vortex-mixed. Diluted 0.2 ml of suspensions were plated onto acidified YMA (1% glucose, 0.5% peptone, 0.3% yeast extract, 0.3% malt extract, 1.5% agar, pH3.5) or Dichloran Rose Bengal Chloramphenicol agar (Merck, Darmstadt, Germany). After incubation at 24uC for 3 days, three (if present) representative colonies of each morphotype were picked for subsequent workup. All yeast isolates were identified by the sequences of the D1/D2 domain of the LSU rRNA genes as described by Lee et al. [40].

Drug susceptibility testing
The MICs of amphotericin B (0.0313-16 mg/L), fluconazole (0.125-64 mg/L), and voriconazole (0.0156-8 mg/L) were determined by the same in vitro antifungal susceptibility testing established in our laboratory [41,42] according to the guidelines of M27-A3 recommended by the Clinical and Laboratory Standards Institute [43]. RPMI medium 1640 (31800-022, Gibco BRL) was used for the dilution and growth of the yeast culture. Strains from American Type Culture Collection (ATCC), including C. albicans (ATCC 90028), C. krusei (ATCC 6258), and C. parapsilosis (ATCC 22019), were used as the standard controls. Growth of each isolate was measured by the Biotrak II plate spectrophotometric reader (Amersham Biosciences, Biochrom Ltd., Cambridge England) after incubation at 35uC for 24 h and 48 h.
The MIC of amphotericin B was defined as the minimum inhibitory concentrations of the drug capable of inhibiting cell growth. For susceptibility to amphotericin B, isolates with MIC §2 mg/L were considered to be resistant and those with MIC, --1 mg/L were susceptible. The MIC for fluconazole and voriconazole was defined as the lowest concentration capable of reducing the turbidity of cells by more than 50%. Isolates with fluconazole MICs$64 mg/L were considered as to have reduced susceptibility and #8 mg/L susceptible. The isolates with MICs in the range of 16-32 mg/L were referred to as susceptible-dose dependent (SDD). Isolates with voriconazole MICs$4 mg/L were considered as with reduced susceptibility and #1 mg/L susceptible. The isolates with MIC 2 mg/L were referred to as SDD.

Pulsed field gel electrophoresis (PFGE)
A total of 18 isolates were examined by PFGE, including 11 soil and 7 human isolates. The soil isolates comprised 6 (NHUE10, 28,33,40,48, and 56) exhibiting fluconazole reduced susceptibility and 5 best matched susceptible ones isolated either from the same sample or from the same type of soil in the same area, NHUE11, 27, 36, 42, and 52 (vs. NHUE10, 28, 33, 40, and 48, respectively). The 7 human isolates, YM990275, 490, 537, 579, 592, 649, and 659, all exhibited reduced susceptibility to fluconazole. Pulsed field gel electrophoresis (PFGE) was performed based on our previous report [44] with some modifications. Briefly, organisms were grown on sabouraud dextrose agar (SDA) plate at 30uC overnight, after which the organisms were embedded in 1% agarose gel (SeaKem Gold agarose, FMC BioProducts). The plugs were subjected to cell lysis by lyticase at 37uC for 2 h, and then treated with proteinase K in 50 mM EDTA buffer at 50uC for 16 h. The genomic DNA was digested with 4 units of BssHII at 37uC for 16 h. Electrophoresis was then performed with a CHEF MAPPER system (Bio-Rad, Hercules, Calif., USA) in 0.56TBE buffer at 6 V/cm at 14uC with alternating pulses at an angle of 120 degrees in a 5-50 s pulse-time gradient for 20 h. A ladder of Salmonella choleraesuis subsp. choleraesuis (Smith) Weldin serotype Braenderup (ATCC BAA664) was used as the molecular weight marker. The PFGE pattern was analyzed by BioNumerics software (Version 4.5, Applied Maths, Kortrijk, Belgium). The position tolerance was set at 1% and optimization was set at 1%. The Dice coefficient was used to analyze the similarities of the band patterns. The unweighted pair group method using arithmetic averages (UPGMA) was used for cluster analysis.

Multilocus sequence typing
A total of 18 isolates exhibiting reduced susceptibility to fluconazole and 5 best matched susceptible ones (either from the same sample or from the same type of soil in the same area), NHUE11, 27, 36, 42, and 52 (vs. NHUE10, 28, 33, 40, and 48, respectively) as well as NHUE 22 with MIC of fluconazole 8 mg/L, were analyzed. Based on our previous report [19], the DNA fragments of six genes: ICL1, MDR1, SAPT2, SAPT4, XYR1, and ZWF1a were sequenced for the analyses. The resulted sequences were aligned with BioNumerics 3.0 (Applied Maths, Kortrijk, Belgium) and compared with those in the database of C. tropicalis (http://pubmlst.org/website) to obtain the identifiers (DST).