Virulence and antifungal susceptibility of microsatellite genotypes of Candida albicans from superficial and deep locations

Abstract A set of 185 strains of Candida albicans from patients with vulvovaginal candidiasis (VVC) and from non‐VVC clinical sources in southwest China was analysed. Strains were subjected to genotyping using CAI microsatellite typing and amplification of an intron‐containing region of the 25S rRNA gene. Microsatellite genotypes of strains from non‐VVC sources showed high polymorphism, whereas those of VVC were dominated by few, closely similar genotypes. However, among non‐VVC strains, two genotypes were particularly prevalent in patients with lung cancer. 25S rDNA genotype A was dominant in VVC sources (86.7%), whereas genotypes A, B, and C were rather evenly distributed among non‐VVC sources; known genotypes D and E were not found. In an experimental mouse model, isolates from lung cancer and AIDS patients proved to have higher virulence than VVC strains. Among 156 mice infected with C. albicans, 19 developed non‐invasive urothelial carcinoma. No correlation could be established between parameters of virulence, source of infection, and incidence of carcinoma. C. albicans strains from VVC were less susceptible to itraconazole than the strains from non‐VVC sources, whereas there was small difference in antifungal susceptibility between different 25S rDNA genotypes of C. albicans tested against amphotericin B, itraconazole, fluconazole, and flucytosine.


Funding information
patients proved to have higher virulence than VVC strains. Among 156 mice infected with C. albicans, 19 developed non-invasive urothelial carcinoma. No correlation could be established between parameters of virulence, source of infection, and incidence of carcinoma. C. albicans strains from VVC were less susceptible to itraconazole than the strains from non-VVC sources, whereas there was small difference in antifungal susceptibility between different 25S rDNA genotypes of C. albicans  (Achkar & Fries, 2010). Mucocutaneous candidiasis can be divided into genitourinary and non-genital disease.
Vulvovaginal candidiasis (VVC) is one of the most frequent manifestations (de Repentigny, Lewandowski, & Jolicoeur, 2004). C. albicans expresses successful adaptation to the human body in response to various stimuli, being able to sustain as a commensal colonizer and as an opportunistic pathogen (Egbe, Paget, Wang, & Ashe, 2015).
Not only the host's immune system is a major factor balancing the transition from commensalism to pathogenicity, but also fungal features as adhesion, hyphal formation, phenotypic switching, extracellular enzyme production, and biofilm formation contribute to C. albicans virulence. Selective proliferation of some genes of commensal strains may precede the transition of C. albicans to infection (Abdulrahim, McManus, Flint, & Coleman, 2013;da Silva-Rocha, Lemos, Svidizisnki, Milan, & Chaves, 2014). The possibility that chronic inflammation, in which C. albicans infection can play a role, may enhance the development of cancer has been brought forward by several authors (e.g., Norgaard, Thomsen, Farkas, Mogensen, & Sorensen, 2013;Ramirez-Garcia et al., 2016).
Epidemiological genotyping is relevant in the context of species prevalence, for elucidation of virulence factors and mechanisms of drug resistance, and is of practical importance to support treatment options in candidiasis. Molecular studies have shown that C. albicans isolates exhibit a high level of genetic diversity. Therefore, reproducible and precise genotyping methods are required, illustrating possible relationships between strain genotypes, phenotypic properties, and host interactions. Strains may differ considerably in virulence, antifungal susceptibility, and other clinically relevant parameters, and appropriate genotyping methods may reveal microevolution by detecting small genetic variations indicating adaptation to long-term host response. In recent years, short tandem repeats or microsatellites have been used as molecular markers for population genetics and genotyping. Microsatellite loci (e.g., CAI, CAIII, CAV, CAVI, and CAVII) located in coding or non-coding regions have been employed for strain typing of C. albicans (Sampaio et al., 2003(Sampaio et al., , 2005. GeneScan analyses of the CAI microsatellites proved to be particularly powerful (Bai, 2014;Carvalho-Pereira, Vaz, Carneiro, Pais, & Sampaio, 2015).
Previous studies showed promising results when using this method to determine genetic diversity in C. albicans with VVC and non-VVC clinical features (Bai, 2014;Fan et al., 2008;Ge et al., 2010). The method proved to be sufficiently sensitive to reveal consistent geographical structuring within the species.
The main aim of our study is to apply genotyping using GeneScan analysis of CAI loci aiming to compare genetic diversities of C. albicans isolates from vaginal candidiasis with those from other clinical sources in patients from the same region. Experimental mouse infection is used to establish comparative virulence of strains from particular patient groups, and in vitro susceptibility is tested to reveal eventual correlations between virulence, source of infection, drug resistance, and genotype.

| Collection and identification of test isolates
A total of 185 independent C. albicans strains were isolated from patients treated in the Affiliated Hospital of Guizhou Medical University (Table S1)

| DNA extraction and polymerase chain reaction
Nuclear DNA was extracted according to methods described by Kaiser, Michaelis, Mitchell, and Laboratory (1994). The rDNA ITS region was amplified using primers ITS1 and ITS4 (Kageyama et al., 2004). Obtained sequence data were compared with those published in GenBank using BLAST. The CAI microsatellite locus was amplified by with primers 5′-ATG CCA TTG AGT GGA ATT GG-3′ (fw) and 5′-AGT GGC TTG TGT TGG GTT TT-3′ (rv) according to Sampaio et al. (2005). For GeneScan analysis, the forward primer was 5′ fluorescently labelled with 6carboxyfluorescein. Polymerase chain reaction (PCR) amplification was performed in a thermocycler (iCycler, Bio-Rad, Hercules, CA, USA) with initial denaturation at 95°C for 5 min, 33 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min and a final extension step of 7 min at 72°C. Fragment sizes of amplified DNA were determined automatically by GeneScan software V3.7. CAI genotypes were designated by the number of trinucleotide repeat units in both alleles (because of the diploid nature of C. albicans) of the microsatellite locus, as described by Sampaio et al. (2005). Primer pairs spanning the transposable intron in the 25S rDNA as follows: CA-INT-L (5′-ATA AGG GAA GTC GGC AAA ATA GAT CCG TAA-3′) and CA-INT-R (5′-CCT TGG CTG TGG TTT CGC TAG ATA GTA GAT-3′). The PCR conditions were denaturation for 3 min at 94°C prior to 30 cycles of 94°C for 1 min, 65°C for 1 min, and 72°C for 2.5 min and a final extension at 72°C for 10 min.

| Antifungal susceptibility
In vitro susceptibilities of Candida strains obtained from blood samples were tested against four antifungal agents: fluconazole (0.12-64 μg/ml),

| Mouse model of disseminated candidiasis
Twelve isolates were used in an experimental mouse model of hematogenously disseminated candidiasis as described previously by Sampaio et al. (2010). Nine of these originated from symptomatic and three from healthy individuals; ATCC 10231 = CBS 6431 was applied as reference. The set included genotypes 21-33 and 21-21, which were primarily derived from patients with lung cancer or AIDS, whereas 30-45, 32-46, and 30-46 were the dominant genotypes of VVC. According to the principle of randomization, control and double-blind strains and saline were divided into 14 groups (12 mice for each group; Table 1). Strains analysed (  Taylor et al., 2000), and 12 mice in the control group were injected with the same volume of saline. All inoculums were confirmed by quantitative culture. For preparation of inoculums, thawed cells from the original stock were grown on PDA medium at 37°C. In each experiment, all isolates were tested simultaneously, and inocula were confirmed by colony-forming unit counting of the suspensions used to infect the mice. Animal welfare was assessed twice daily during 5 months. The mice were monitored at least two times daily, and dead mice were dissected immediately. Brain, liver, spleen, and kidneys from each mouse were harvested and weighed. After 5 months of infection, all mice were sacrificed for anatomical examination.

| Histopathology
Organs excised from all infected mice were fixed in 10% phosphatebuffered formalin, embedded in paraffin, sectioned, and stained with haematoxylin-eosin (H&E) according to Kumar, Saraswat, Tati, and Edgerton (2015). On the basis of 1,000 neoplastic nuclei, Ki-67 labelling index was calculated in each slide as the percentage of immunopositive nuclei (Šteňo et al., 2014).

| Statistics
Survival of mice was compared with log-rank test and Gehan-Breslow-Wilcoxon test, using GraphPad Prism 7 for Windows. P values ≤.05 were considered to be significant. 25S rDNA group I intron genotypes for C. albicans from different sources and carcinoma of mice in each group were compared using chi square (Li, 2010  Overall, all of the strains were susceptible to amphotericin B. There

| Murine model
Five-month survival rates of all mice are listed in Table 4. Using the sources of isolation as parameter, strains could be divided in lung cancer, HIV/AIDS, VVC, and asymptomatic individuals. On the basis of strain sources and analysed by log-rank test and Gehan-Breslow-Wilcoxon test, overall differences in survival were highly significant (p < .05). Isolates from lung cancer and AIDS proved to have higher virulence than VVC strains (Figure 2). Figure 3a shows histopathology of kidney derived from mice died within 48 days. Flat urothelial hyperplasia comprised an increased number of cell layers with few or no significant cytological abnormalities; fungal cells were predominantly in the hyphal form. Polarity was preserved, and none to minimal variation in architectural and nuclear features was noted. 8130/23; John et al., 2004). It is important to note that linear expression with occasional local interruption was observed. Table 1 shows carcinoma occurrence of mice infected with C. albicans strains.
Abnormal cells were observed on pathologic sections of murine   (Figure 4a). Figure 4b shows C. albicans infection after 48 days; the Ki67 labelling index was 40%, leading to a diagnosis of non-invasive urothelial carcinoma, high grade.

| DISCUSSION
In the current study, genotyping of C. albicans strains from VVC and non-VVC disorders using microsatellite analysis of CAI and 25S rDNA revealed that most VVC strains possessed specific genotypes.
Logistic regression analysis showed that, compared with other genotypes, the prevalent CAI genotypes of VVC have a high risk of OR (OR = 4.3). It shows that the risk of VVC in prevalent genotype was 4.3 times higher than that in other genotypes after controlling for age, marital, and sexual factors. CAI microsatellite genotyping has repeatedly been proven to be a powerful method for epidemiology of C. albicans (Bai, 2014). This approach revealed a significantly larger diversity than classical typing using 25S rDNA introns. The Out of 57 microsatellite genotypes revealed in this study, 12 seemed to have a global distribution, because the same genotypes were reported in Europe (Kageyama et al., 2004). Thirty-six genotypes were shared between Beijing/Tianjing, Guangdong, Guiyang, and Wuhan, representing northern, southern, western, and central regions of China, respectively (Bai, 2014;Fan et al., 2008;Ge et al., 2010).
Seven novel genotypes were recognized in the current study, which seemed to have a local distribution.
Genotype CAI microsatellite patterns were also structured by association with clinical types. Profiles of systemic infections were found to differ significantly from those of VVC. The majority of randomly taken C. albicans strains associated with VVC belonged to the same or highly similar CAI genotypes. Most isolates from patients with lung cancer were recognized as CAI 21-33 and similar genotypes. In contrast, the overall pattern of strains originating from respiratory diseases and from AIDS patients was highly diverse. In respiratory disease, the diversity may be explained by local severe abuse of antibiotics to treat respiratory disorders, which has a negative impact upon the host's immune system and thus predisposes to fungal infection; the number of opportunistic infections in those patients is predicted to be high. In addition, the use of sputum samples rather than bronchoalveolar lavage fluid or blood may have led to the inclusion  CAI types occurring in 23 strains, also seem to be infected easily by diverse C. albicans genotypes (Figure 1 and Table S1).
rDNA genotypes A and B versus C differ by absence and presence of a 25S rDNA-based group I intron. C. albicans strains with relatively effective adherence properties are known to have higher virulence than non-Albicans Candida species such as Candida krusei (Ferreira, Prado, Carvalho, Dias, & Dias, 2013). Genotype A seems to be more virulent than genotypes B and C (Sardi, Duque, Hofling, & Goncalves, 2012) and is dominant in VVC and in healthy oral mucosa of all age groups (Xu, Chen, & Li, 2008). In our study, all strains of the preponderant microsatellite genotypes from VVC (mainly 30-45, 32-46, and 30-46) belong to rRNA genotype A, which is consistent with current literature.
The hypothesis that strains differing in predilection for body sites might differ in virulence has been verified with animal inoculation experiments (Li et al., 2015). A progressive reduction of virulence was observed, with systemic isolates being more virulent than VVC strains (Figure 2), judged from the time of death after challenge and from pathologic expression in the infected mice. We also showed differences in pathologic anatomy of various tissues after prolonged survival of Candida yeast cells. The Ki-67 labelling indices of proliferating cells in mice that died after 3 weeks were mostly around 3%. A threshold Ki-67 labelling index of >3% immune reactivity is termed "atypical adenoma," which suggests a potential stimulus by Candida cells (Heaney, 2011a(Heaney, , 2011bŠteňo et al., 2014).
In mice died 2 months after inoculation, obviously non-invasive urothelial carcinoma was observed (Dalgliesh et al., 2010), a disease associated with tuberous sclerosis that can result in kidney cancer (Kapoor, Sikri, Grover, Malhotra, & Sachdeva, 2014;Ramirez-Garcia et al., 2016). A recent study supports the view that C. albicans infection may not just be a consequence of cancer but is an actively con-  (Krogh, 1990).
In addition, alcohol dehydrogenase (ADH1) produced by C. albicans The degree of damage to kidneys in inoculated animals was found to be more severe than to other organs, irrespective of genotype or clinical origin of the inoculated strain. Earlier publications also mentioned that Candida infection was less controlled in the kidney than in other organs (Li et al., 2015). Yeast cells penetrated through blood vessels into both cortex and medulla, causing an influx of neutrophils. Several authors mentioned an association between inflammation and cancer (Balkwill & Mantovani, 2001;Coussens & Werb, 2002). In the present experiment, non-invasive urothelial carcinoma (high grade) was found in the experiment mice population. The mechanism that caused the appearance of this phenomenon is as yet unclear but requires close attention. We emphasize that there is a need not only to control Candida infection during cancer therapy, especially given the important role of this yeast in nosocomial infections, but also to find new therapeutics to suppress protumour effects by C. albicans.
Strains with rDNA genotype A have been shown to have decreased susceptibility to the antifungal agent flucytosine (McCullough, Clemons, & Stevens, 1999). Heterogeneity in resistance between strains is known to exist, but this was ascribed to geographic variation (Fan et al., 2008). Our strains of genotype A originating from VVC showed increased resistance to itraconazole but neither to flucytosine nor to other drugs tested. The correlation of the absence of a group I intron in the 25S rRNA with high levels of resistance to flucytosine was therefore not confirmed. The found deviations in drug resistance profiles of C. albicans genotype A may thus have been spread by local epidemics. The result of antifungal susceptibility tests showed that all of C. albicans strains in our study were susceptible to amphotericin B and there was a small difference in susceptibility to flucytosine and fluconazole; however, the strains from VVC were less susceptible to itraconazole than strains from non-VVC infections.
For some clinicians, the azoles have become the drug of choice for treatment of Candida infections, because of high bioavailability, extensive tissue penetration, long half-life, convenient use, and good safety.
Especially, the superficial infection VVC is more diverse than deep infection for local application. But the widespread overuse induces the emergence of drug-resistant strains, resulting in a decline in susceptibility, when clinical medications do not follow the standard guidelines.