A Predominance of Clade 17 Candida albicans Isolated From Hemocultures in a Tertiary Care Hospital in Thailand

Candida albicans is one of the most common human fungal pathogens. Candidemia has significant mortality globally. No epidemiological study of C. albicans based on multilocus sequence typing (MLST) has been conducted in Thailand. Therefore, MLST was used to study the molecular epidemiology of C. albicans blood strains in a large Thai teaching hospital. In vitro virulence phenotypes and antifungal susceptibility testing by broth microdilution were also conducted. Forty-six C. albicans blood strains from 37 patients were collected from the Department of Microbiology, Siriraj Hospital, in 2016 and 2017. Most patients (71.8%) were more than 60 years old, and the case fatality rate was 54.8%. The male-to-female ratio was 5:3. Thirty-four diploid sequence types (DSTs), including six new DSTs, were identified, with DST2514 (8.7%) and DST2876 (8.7%) as the most common DSTs. Strains were clustered into nine clades. Unlike other studies of C. albicans blood strains in Asia, clade 17 was the most common (13 strains, 28.3%). Sequential allelic changes were evident in sequential strains from one patient. All strains produced phospholipase and hemolysin, while none produced proteinase. The ability to form biofilm was found in 82.6% of the strains. Clade 17 strains showed significantly stronger hemolytic activity than non–clade 17 strains (69.2% versus 27.3%; p = 0.022). However, no significant association existed between clades and patient mortalities. All were susceptible or wild type to anidulafungin (MIC range = 0.015–0.12 and GM = 0.030), micafungin (MIC range = ≤ 0.008–0.015 and GM = 0.008), caspofungin (MIC range = 0.008–0.12 and GM = 0.036), and amphotericin B (MIC range = 0.25–0.5 and GM = 0.381). Only one strain was resistant to voriconazole (MIC range = ≤ 0.008 to ≥ 8 and GM = 0.010) and fluconazole (MIC range = 0.12–16 and GM = 0.398). In conclusion, a high prevalence of clade 17 C. albicans blood strains was found in Thailand, in contrast to other Asian countries. This unique finding might be explained by the strong hemolytic activity that is required for bloodstream infection of C. albicans.


INTRODUCTION
Candida albicans is a member of the human flora and is commonly colonized in the human digestive tract. C. albicans is an important cause of candidiasis (Bougnoux et al., 2006). It can cause infections that range from superficial infections to life-threatening systemic infections. Approximately 75% of all women suffer from vulvovaginal candidiasis at one point in their life span, and C. albicans accounts for nearly 90% of such cases (Yazdanparast et al., 2015). C. albicans and other Candida species including C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei can travel through the bloodstream and together are causative agents in 90% of invasive candidiasis cases . Many risk factors contributing to candidiasis were reported, including host-related factors such as immunosuppressive disorders, neutropenia, age, or debilitating diseases, and healthcare-associated factors such as catheter use, surgical interventions, transplantation, or the use of antimicrobial drugs (Yapar, 2014;Diba et al., 2018).
Virulence is defined as the ability of a microbe to damage its host (Casadevall and Pirofski, 1999), and C. albicans possesses a wide range of virulence factors, including phenotype switching, adhesin expression, thigmotropism, biofilm formation, and extracellular hydrolytic enzyme secretion, to facilitate tissue invasion and ultimately produce the diseases (Mayer et al., 2013). A recent trend showed increased resistance of non-albicans Candida to azole-based antifungal agents, echinocandins, and multiple classes of drugs (Wiederhold, 2017). However, C. albicans remains highly susceptible to most antifungal drugs (Bustamante et al., 2014;Posteraro et al., 2015;Doi et al., 2016;Tadec et al., 2016;Tan et al., 2016;Chapman et al., 2017).
So far, no epidemiological study of C. albicans based on MLST has been conducted in Thailand. Therefore, MLST was used to study the molecular epidemiology of C. albicans strains of candidemia patients at Siriraj Hospital, including virulence factors and antifungal susceptibility profiles.

Study Sites and Subject Selection
All C. albicans strains were collected from positive hemocultures from a diagnostic microbiology laboratory of the Department of Microbiology, Siriraj Hospital, between April 2016 and November 2017. This study was approved by the Siriraj Institutional Review Board, no. SI 091/2016. Informed consent was not required in this work. All strains, collected from the same patients, were considered sequentially collected, as blood samples were taken from different body sites at different time points. Forty-six strains from 37 individual candidemia patients were collected. Available patient records were retrieved from the hospital registry. A healthcare-associated infection was defined as an onset of candidemia symptoms and/or signs after 48 h of admission. As the hospital is a tertiary care referral center, patient origins varied. Most cases (81.25%; 26 of the 32 available patient records) were from central Thailand; 6.25, 6.25, and 6.25% of cases were from Eastern, Northeastern, and Western Thailand, respectively. The case fatality rate was calculated as the ratio of patient deaths to all candidemia cases in this study. All strains were identified using CHROMagar Candida chromogenic media (Oxoid, United Kingdom) and RapID TM YEAST PLUS System (Thermo Scientific, United States) followed by singlecolony subculture on Sabouraud dextrose agar (SDA; Oxoid, United Kingdom) to prepare pure cultures before performing any further analysis.

MLST Analysis
The seven housekeeping genes including AAT1a, ACC1, ADP1, MPIb, SYA1, VPS13, and ZWF1b were used for MLST according to a previously published method (Bougnoux et al., 2002). Briefly, PCR mixtures were prepared in 50 µl reaction volumes containing DNase-RNase free water, 10 × buffer, 25 mM MgCl 2 , 2.5 mM dNTPs, 10 µM for each of the forward and reverse primers of each gene, 0.25 µl of 5 U/µl Taq DNA polymerase (Thermo Scientific, United States), and genomic DNA. The condition was set up with a denaturation step at 95 • C for 5 min, followed by 35 cycles of 95 • C for 40 s, annealing at 50 • C for 40 s, extension at 72 • C for 1 min, and a final extension step at 72 • C for 4 min. Approximately 50 ng of DNA was sequenced using each forward and reverse primer on a DNA analyzer (First Base Company, Singapore). The generated sequences were manually edited using MEGA6 software and aligned using Clustal W. The allelic profile (allele number) for each gene and allele combination (diploid sequence type, DST) for the seven loci of each isolate was assigned, or new DST numbers were obtained from the C. albicans MLST database 1 . For cluster analysis, because the base at each polymorphic site can be homozygous or heterozygous, the nucleotide bases were manually edited according to a previous publication (Tavanti et al., 2005). The bases of each taxon were doubled when the site was homozygous and written in pairs when the site was heterozygous. Relationships among concatenated sequences of the seven loci of each strain were determined using cluster analysis using the unweighted pair group method using their arithmetic averages (UPGMA) and the neighbor joining (NJ) method determined by p distance with pairwise deletion of the MEGA 6 software. A bootstrap of 1,000 replications was used for the construction (Shin et al., 2011;Wu et al., 2017). Bootstrap values of ≥75% were defined as statistically significant. Finally, clade numbers were assigned according to previous publications (Odds et al., 2007;Tsai et al., 2015;Wu et al., 2015;Gong et al., 2018) or by using the eBURST method as described previously (Odds et al., 2007). DSTs from clade numbers that were not presented in our strains were also included as reference DSTs for UPGMA and NJ analysis (Figure 1).

In vitro Virulence Study
The test for C. albicans phospholipase activity was performed on egg yolk medium according to a previous study FIGURE 1 | Genetic diversity inferred by UPGMA. Bootstrap support values above 75 for UPGMA (bold) and NJ method (italic bold) were indicated. Ref; reference DST, pt; patient number. (Price et al., 1982). Strains were grown on SDA for 24 h before adjusting to a 0.5 McFarland standard (10 8 cells/ml). One hundred milliliters of medium containing 13.0 g SDA, 11.7 g NaCl (Merck, Germany) SDA, 11.7 g NaCl (Merck, Germany), 1.48 g CaCl 2 .2H 2 O (Merck, Germany), 2% agar (Oxoid, United Kingdom), and 10% sterile egg yolk (Merck, Germany) was prepared to determine extracellular phospholipase activity (Tsang et al., 2007). Five microliters of a yeast suspension (10 8 cells/ml) of the tested and control strains was dropped onto the surface of the egg yolk-containing plate and incubated at 37 • C for 48 h to measure a precipitation zone around the colony. Extracellular phospholipase activity of C. albicans was considered positive when a white precipitation zone was visible around the colony on the plate (Macfarlane and Knight, 1941).
Candida albicans strains' proteinase activity was analyzed on bovine serum albumin medium by the modified Staib method (Staib, 1965;Sachin et al., 2012). Five microliters of a yeast suspension (10 8 cells/ml) of the tested and control strains was dropped on a medium containing 1.17% yeast carbon base (Becton Dickinson, United States), 0.2% bovine serum albumin (GoldBio, United States), 2% agar (Oxoid, United Kingdom), and 0.01% yeast extract (HiMedia, India) and then incubated at 37 • C for 2 days. The presence of a clear halo zone around the colony was recorded as evidence of proteinase activity.
To measure hemolytic activity, a medium was prepared by adding 7 ml fresh sheep blood (Clinag, Thailand) to 100 ml SDA supplemented with 3% of enriched glucose (HiMedia, India) (Luo et al., 2001). Five microliters of a yeast suspension (10 8 cells/ml) of the tested and control strains was spotted onto the medium. The plate was then incubated at 37 • C for 48 h. The presence of a distinct translucent halo indicates positive hemolytic activity.
The ratios of the diameter of the yeast colony to the total diameter of the colony plus the white precipitation zone, the clear halo zone, and the translucent halo zone were used to represent the level of phospholipase, proteinase, and hemolytic activity, respectively (Sachin et al., 2012).
Biofilm formation was determined using the 2,3bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carbox anilide (XTT; Thermo Scientific, United States) reduction assay using 96-well microplates as indicated in previous studies (Yigit et al., 2011;Oz et al., 2012). Briefly, yeast cells were adjusted to a 0.5 McFarland standard with yeast extract-peptone-dextrose medium. One hundred microliters of yeast suspension was seeded in 96-well plates and incubated at 37 • C for 48 h. The cells were then washed twice with phosphate-buffered saline to remove non-adhering cells. A 50-µl mixture of 1 mg/ml XTT and 10 µl of activation reagent, phenazine methosulfate (HiMedia, India), was added to the wells and further incubated in the dark for 2 h. The biofilms were quantified at A 490 and defined as low biofilm formation and high biofilm formation by comparing the XTT activity of each strain to the geometrical mean as described previously (Li et al., 2003;Hasan et al., 2009).

Statistical Analysis
PASW Statistic 18 (IBM Corporation Armonk, NY, United States) was used for data analysis. To estimate frequencies, descriptive statistical tools were used. GM for biofilm formation was calculated using Microsoft Excel 2013 (Microsoft Corporation, Redmond, WA, United States). Statistical significance of an association between the MLST clusters and virulence phenotype was determined using chi-square or Fisher's exact test (PASW version 18, IBM, Bangkok, Thailand); p ≤ 0.05 was considered statistically significant.

Most Candidemia Cases Occurred as Hospital-Associated Infections and in Older Patients
Clinical data were available from 32 of the 37 patients. All patients (100%) had one or more immunocompromised conditions, including receiving immunosuppressants or chemotherapy (10 patients, 31.3%), having diabetes mellitus (10 patients, 31.3%), being a newborn (2 patients, 6.2%), and old age (>60 years old) (23 patients, 71.8%). The case fatality rate was 54.8%. The maleto-female ratio was 5:3 (Supplementary Table S1). Most cases (31 patients, 96.9%) were hospital-associated infections except one who had Candida sepsis as a primary diagnosis. A primary source of candidemia was not found in patients in this study.

Sequential Blood Strains Showed Evidence of Sequential Allelic Changes
Five of the 37 patients had sequential strains isolated from their blood. Interestingly, three of these five patients were infected with more than one DST at the same time. Two patients had mixed infections with two DSTs from blood cultures collected on the same date (patient 12 and patient 34; Figure 1 and Supplementary Table S1). Furthermore, C. albicans isolates from patient 8 had three different DSTs identified from sequential blood cultures on days 0, 3, and 5 as DST1710, DST2924, and DST2527, respectively (Supplementary Table S1). Further analysis of diploid allele types (DATs) of the three sequential strains showed sequential changes of only one DAT at a time of the two later-collected strains (Table 1). Specifically, DAT7 of the VPS13 gene of the S29 strain collected on day 0 changed to DAT100 of the S30 strain collected on day 3. Then, DAT5 of the ADP1 gene of the S29 strain changed to DAT6 of the S1 strain collected on day 5. The profiles of the allelic genes from the patients in whom these

Clade 17 Strains Showed Significantly Stronger Hemolytic Activity
As the clade 17 predominance in C. albicans blood strains is uncommon, the virulence phenotype and antifungal susceptibility testing results of clade 17 and the other clades were determined and compared. Four virulence phenotypes were tested. First, the phospholipase zone values of the C. albicans strains ranged from 0.50 to 0.89. All C. albicans strains were phospholipase producers ( Table 2 and Supplementary Table S1).
Most C. albicans strains showed very strong phospholipase activity (30 strains, 65.2%). No weak phospholipase producer was observed. The mean phospholipase activity of all strains was 0.675 (±0.072). Second, none of the C. albicans produced proteinase (Supplementary Table S1). Third, hemolytic activity was detected in all strains ( Table 2 and Supplementary Table S1).
The hemolytic zone values ranged from 0.50 to 0.95. Most strains had positive activity (60.9%). The mean hemolytic activity of the C. albicans strains was 0.666 (±0.089). Finally, most C. albicans strains (54.3%) were highly able to form a biofilm layer on the bottom surface of each well ( Table 2 and Supplementary  Table S2). Association analysis using the chi-square test revealed that clade 17 strains showed significantly stronger hemolytic activity than non-clade 17 strains (69.2% versus 37.5% and 24%; p = 0.025). There was no significant association between the MLST clades and other virulence phenotypes (p > 0.05) ( Table 2).

Clade 17 Showed No Significant Difference in Antifungal Susceptibility Pattern
All strains were susceptible to anidulafungin, micafungin, and caspofungin, and wild type to amphotericin B. Only one isolate (2.2%) was resistant to both voriconazole and fluconazole. The MIC50, MIC90, and geometric mean of all nine drugs are shown in Table 3. There was no significant correlation between clade and antifungal susceptibility. Further association analysis revealed that the clade 17 strains did not cause significantly higher case fatality rates than clade 1 and the other clades (75% versus 42.9% and 55%; p = 0.340) (Supplementary Table S2).

DISCUSSION
In developing countries, candidemia case fatality rates of more than 50% have been reported, which is similar to    the 54.8% case fatality rate found in this study (Kaur and Chakrabarti, 2017). Antibiotic resistance is generally one of the common causes of treatment failure of bloodstream infections that leads to patient fatality. However, such resistance is very rare for C. albicans blood strains. A comprehensive 2016 study of C. albicans bloodstream infections in East and Southeast Asia in found that just 1 out of 309 C. albicans blood strains resisted fluconazole, and all strains were sensitive to echinocandins (Tan et al., 2016). Similarly, C. albicans blood strains in our study were also highly susceptible to both echinocandins and azoles. This suggests that antifungal resistance is not a cause of treatment failure for C. albicans bloodstream infections. Rather, candidemia mortality is associated with patients' underlying conditions or management such as neutropenia from chemotherapy, septic shock, or ICU admission (Boonyasiri et al., 2013). However, comparisons of these factors for C. albicans and non-albicans Candida are lacking. Candidemia risk factors are either healthcare related or host related (Kaur and Chakrabarti, 2017). In this study, almost all patients had both types of risk factors, specifically healthcare-related infections and immunocompromising conditions. In fact, none of the 32 patients whose clinical data were available had no risk factors. However, indepth details of the candidemia risk factors, including specific healthcare-related factors (such as use of a central vascular catheter, total parenteral nutrition, urinary catheter) and previous colonization of C. albicans (Kaur and Chakrabarti, 2017), were not available in our database. With the absence of primary sources regarding candidemia in Thailand, as evidenced by this study (Boonyasiri et al., 2013), determining C. albicans colonization is crucial, but it is not common practice in Thailand. Further collection and comprehensive analysis of such data would give an insight into specific risk factors for candidemia in Thailand.
High diversity among sequence types of C. albicans was reported in several previous studies (Odds et al., 2007;Shin et al., 2011;Alastruey-Izquierdo et al., 2013;Asadzadeh et al., 2017). Among 46 clinical strains in this study, high sequence variation was also observed with 34 distinct DSTs. DST254 was derived from the parental DST69, the most common DST found using MLST worldwide (Odds et al., 2007). However, DST69 was not found in this study. Additionally, DST443 and DST693, found in 6.5 and 2.2% of cases in this study, respectively, were the most prevalent DSTs in Taiwan (4.6 and 6.7%, respectively) and China (DST443, 7.7%) (Hu et al., 2015;Wang et al., 2015). DST443, DST457, and DST445 are particular sequence types found in Asia  that were also present in this study.
In general, clade 17 is more common in Asia (8.2%) than other continents (0-6.2%) (Odds et al., 2007). Moreover, a 2015 Chinese study also reported clade 17 as the second most common clade identified from 40 C. albicans non-blood strains (Wu et al., 2015). However, when considering only blood strains, clade 17 was rare (3.8%) in the globally collected C. albicans samples, and clade 1 was the most common (27.8%) (Odds et al., 2007). Furthermore, no clade 17 strains were found in 149 C. albicans blood strains collected in a recent study from Korea (Jung et al., 2016). Only 10.9 and 15% of C. albicans blood strains were identified as clade 17 in Taiwan  and China (Wu et al., 2015), respectively. Therefore, virulence studies were conducted to determine whether the Thai clade 17 isolates were more virulent than isolates from the other clades. The unusually high clade 17 prevalence of the Thai C. albicans blood strains might be explained by the stronger hemolytic activity that is required for C. albicans bloodstream infection (Furlaneto et al., 2018). However, clade 17 strains did not cause a significantly higher case fatality rate than the other strains. This suggests that many other factors also affect pathogenesis and patient mortality, especially host and environmental factors (Boonyasiri et al., 2013;Kaur and Chakrabarti, 2017).
Phospholipase enzyme plays an important role in host tissue invasion by disrupting the epithelial cell membrane.
In general, C. albicans strains isolated from blood samples had higher enzyme activity than other anatomical sites (Ibrahim et al., 1995;Mohan Das and Ballal, 2008). For example, blood strains of C. albicans from Saudi Arabia (Fotedar and Al-Hedaithy, 2005), the European SENTRY program (Borst and Fluit, 2003), Brazil (Mattei et al., 2013), and Turkey (Atalay et al., 2015) showed 100, 71, 78, and 88% positive phospholipase activity, respectively. Similarly, all C. albicans in this study showed positive phospholipase activity. However, despite their similar function regarding host invasion, unlike phospholipase, proteinase activity was typically undetectable in blood strains, whereas 95% of the strains from other sources had proteinase activity (Kantarcioglu and Yucel, 2002). In this study, proteinase activity was undetected in all C. albicans strains. Conversely, the European SENTRY program reported that approximately 80% of strains were positive for proteinase (Borst and Fluit, 2003). This suggests that proteinase activity variations might depend not only on the infection source but also on the strains' geographic location (De Bernardis et al., 2001).
Both hemolysin and biofilm formation are important virulence factors for C. albicans (Mayer et al., 2013). Hemolysin facilitates iron acquisition in disseminated candidiasis (Tsang et al., 2007), while biofilm formation protects the yeast from a harsh environment including nutrient storage, metabolic cooperation, and acquisition of new genetic traits (Mayer et al., 2013). In our study, C. albicans highly expressed both hemolytic activity (100%) and biofilm formation (82.6%). This suggests that both are important virulence factors for bloodstream infection.
Globally, C. albicans showed high susceptibility toward fluconazole, voriconazole, and echinocandins. Azole resistance was generally region dependent and uncommonly found in C. albicans (Lyon et al., 2010;Tan et al., 2016). All C. albicans strains in this study showed high susceptibility to anidulafungin, micafungin, and caspofungin, similar to previous studies from Australia, the Asia-Pacific, and Italy (Prigitano et al., 2016;Tan et al., 2016;Chapman et al., 2017). Only one (2.17%) C. albicans isolate was resistant to voriconazole and fluconazole. This isolate was identified as DST2527, which resisted both voriconazole and fluconazole. This finding contrasts with another study from Taiwan . Therefore, association of DST2527 with resistance is unlikely. Moreover, when categorizing strains based on ECVs, all C. albicans strains also had wild-type MICs for amphotericin B, similar to previous studies (Petlum et al., 2013;Posteraro et al., 2015;Tan et al., 2016;Chapman et al., 2017). These data confirm echinocandins, fluconazole, and amphotericin B as the drugs of choice for treating patients with candidemia (Pappas et al., 2016). Moreover, a recent study showed that a combination of echinocandin and triazole drugs had potent activity for Candida spp., and triazole drugs had potent activity for Candida spp. . Therefore, antifungal testing for such a combination drug is beneficial, and its interpretation criteria should be established.

CONCLUSION
In conclusion, the molecular epidemiology among C. albicans blood strains in Thailand showed high genetic diversity. No significant association existed between patient fatality rate and the MLST clades. Clade 17 strains as a predominant clade among isolates showed significant association with hemolytic activity. This finding requires further strain sampling and pathogenetic study.

ETHICS STATEMENT
This study was approved by the Siriraj Institutional Review Board, No. SI 091/2016. Informed consent was not required in this work.

AUTHOR CONTRIBUTIONS
PN, PC, and KS designed the study. LP and SP performed the experiments and analyzed the data. LP, SP, and PN wrote the manuscript. All authors read and approved the manuscript.

FUNDING
The study was supported by the Siriraj Research Development Fund, grant number (IO) R016033021, Faculty of Medicine Siriraj Hospital, Mahidol University to PN, and Siriraj Graduate Scholarship to LP.