Genetic Structure of Candida glabrata Isolates from Hospitalized and Non-Hospitalized Patients in Sfax-Tunisia

The use of molecular technique to determine genotype of clinical isolates can provide useful clues for estimating the reproductive mode, geographic prevalence’s, and cross transmission of bloodstream infections between patients [712]. Possible transmission between patients has been suggested and clusters of invasive infections have been reported [13-15]. Molecular typing method would allow better understanding of the emergence of this species, notably in a hospital context. In the present study, we aimed to analyze by polymorphic markers a collection of C. Glabrata isolates from various sites in hospitalized and non hospitalized patients to look for association of variability with patient population.


Introduction
During the last decade, incidence of invasive candidiasis with non-albicans Candida species has increased [1,2].Candida glabrata is the most significant species that has emerged and now regularly ranks number two, after C. Albicans, as the etiologic agent of superficial and invasive candidiasis occurring in adults [3,4].Reasons for this change in species distribution remain uncertain but may be partially due to the natural resistance of

Statistical analyses
The genetic differentiation between both populations was measured using Wright's Fixation Index (FST) computed with the Fstat V2.9.3 software developed by Goudet J. (http://www2.unil.ch/popgen/softwares/Fstat293dist.exe).FST is a measure of genetic divergence among populations [20].The Fst value = 0.05 indicates little genetic divergence between populations while moderate genetic differentiation was considered as 0.05 < Fst < 0.15.A great genetic differentiation is defined as FST value ranged from 0.15 to 0.25.
Nei's genetic diversity was used as measure of gene diversity and differentiation among the populations [21].To determine the extent of clonality and recombination in the different populations, the Index of Association (IA), rbard, rbars and Proportion of compatible pairs of loci were computed with Multilocus 1.3b software using 1000 randomizations tests to obtain reliable p-values and without considering repeated genotypes [22].All test the null hypothesis of no linkage disequilibrium, which would indicate the absence of recombination.They are expected to be zero if populations are significantly freely recombining and greater than zero if there is association between alleles (clonality).The rbard statistic takes into consideration the number of loci tested and is considered a more robust measure of association and rbars test whether alleles tending in the same direction are positively or negatively associated.Prcompat should be 1 if all Phylogenetic relationships were estimated based on multilocus genotypes using the Neighbor-Joining method as implemented in http://pubmlst.org/.The effect of covariates on the cluster distribution was tested with the Chi-squared test.

Genotyping analysis
In the present study, a total of 145 Candida glabrata isolates from inpatients and outpatients were typed using polymorphic markers (MTI, ERG3, GLM4 and GLM5).From hospitalized and non hospitalized patients, three and two separated subgroups according to site and time of isolation were characterized respectively (Table 2).Multilocus analyses resulted in the delineation of 62 genotypes (Table 2).Forty eight of the multilocus genotypes, corresponding to 33.1% of the total number of isolates, were represented by single isolate.2) in addition to 30 alleles and 37 genotypes previously described [14].

C. Glabrata genotyping were performed by only four
Overall, multilocus analysis of the 145 isolates showed that genotypes G10, G26, G9, G13 and G22 were the most frequent and represented respectively 18.6%, 14.48%, 7.58%, 5.51% and 4.8%.These multilocus genotypes together represented 51% of the entire population and were identified in all studied isolation sites except buccal site which has been characterized by unique genotypes (G41, G43, G51, G53-G60).These five genotypes represented 72.7% of the invasive candidiasis population When comparing isolates from inside and outside hospital, we noted that each population was dominated by the same genotypes and shared alleles and frequency in most cases.The frequency of some genotypes varied from one population to another although it was not significantly associated (p ≥ 0.05), i.e G22 was more common in invasive population (5.1%) and completely absent in outside of the hospital.

Population genetic analysis
In order to assess differentiation among populations, Wright Fixation Index (Fst) was calculated between hospitalized and non hospitalized population (Fst = 0.0274).This result indicated a little genetic differentiation suggesting a non specific genetic structure of C. Glabrata population from inside or outside the hospital.
Genetic diversity varied from 0.85 for vaginal population to 0.95 in invasive isolates (2009-2011) and buccal samples (ID = 1) (Table 3).Buccal populations were considered the most divergent followed by invasive candidemia (2009-2011).In addition, there were a higher number of strains with a unique genotype within 13 among 13 isolates of buccal population than invasives isolates (2009-2011) ( Neighbor-joining tree was constructed for the entire population based on the polymorphism of the four polymorphic markers as shown in Figure 1.There was no unique topology for intra or extra-hospital isolates.Isolates from different anatomic location were subdivised into different genotypes (Figure 1).
A neighbor-joining tree was constructed to show genetic relationship among 62 observed multilocus types in our combined population (Figure 2).Our population was subdivided into four major groups.The first group contains a unique multilocus type G30 from urinary population, the second contain G8 from vaginal population, the third contain the G10 which was the most frequent genotype and the fourth contains all other types including four of the major multilocus types G10, G26, G13 and G22.
To determine the extent of clonality and recombination in different populations, we used four measures of associations (IA, rbard, rbars and Proportion of compatible pairs of loci) [18].For all studied sub-populations, a randomization test for linkage disequilibrium measures presumably rejected the null hypothesis of recombination as well as the total isolates considered to be a single population although proportion of compatible pairs of loci reveals some degree of recombination (prcompat ˂ 1) (Table 1).
The antifungal susceptibility to fluconazole of these strains was not different between hospitalized and non hospitalized patients (Table 1).Among 59 invasive C. Glabrata isolates and 86 non invasive isolates, only 15 isolates were resistant to fluconazole (10.34%).Genotypic analysis of these isolates revealed that they had different genotypes and were subdivided into 8 multilocus types (Figure 1) and the resistant phenotype was not significantly associated with a population or another (p ≥ 0.05).

Discussion
We studied a genetic population structure of C. Glabrata isolates from inside and outside of the CHU hospital in Sfax (Tunisia).Polymorphic markers analysis revealed that all loci showed a high degree of diversity and notably each locus represented 6 -11 alleles (Table 2).Most of them were common to the majority of isolates but few were less common and were identified in only one or at least two isolates.
Genotypic distribution remains similar to that described in Tunisia in previous study when analyzing 85 unrelated isolates from different anatomical sites by six microsatellites [18].In Tunisian population, the same genotypes (G10, G26 and G9) dominated.G10 was the most frequent type which was observed in 27 clinical isolates.The latter type was identified in phylogenetic tree as a single branch which seems to be stable and did not undergo natural evolution process.The overall abundance of G10 but also G26 and G9 suggest that at least some isolates with identical types may be clonally related by descent.
In addition, we have also provided that there was a significant clonal component for the entire population (IA = 0.198).
In our report, we revealed that the inpatient population was not significantly differentiated from out-patient population (Fst <    0.05).Moreover, little differentiation was conducted for all pair wise population except buccal population which was considered moderately differentiated especially from invasive candidemia population.This latter result may be due to the specific virulence and resistance capacities of C. Glabrata in oral mucosa and digestive tract isolates [23,24].
Despite a lack of differentiation between the two populations (Fst = 0.0274), each of them included some genotypes which were not found in the other (Figure 1).Indeed, the inside hospital population was more divergent and showed the inclusion of 38 multilocus types.On the other hand, the outside hospital population was less divergent but showed the inclusion of 18 multilocus types not described in inside hospital.
In our setting, when analyzing 108 isolates from hospitalized patients and 37 from non hospitalized patients, we identified one multilocus type for 2.5 isolates for hospital environment and one multilocus type for 1.6 isolates in outpatients.Other studies, when analyzing intra-population diversity by Multilocus Sequence Type (MLST) established one multilocus type for every 4.8 patients in Baltimone in 2008 and one multilocus type for every 2.4 isolates in Atlanta from 1992 to 1993 [10,11,25].These results demonstrated that our method may be more discriminative in examining population structure.
Most of C. glabrata fungemia are thought to originate from the endogenous flora of the host when host factors or environmental conditions allow them to switch from a commensal state to a pathologic state.In order to predominate as a pathogen, an isolate must first predominate as a commensal.In our study, three genotypes G10, G26 and G9 predominated in both invasive and non-invasive population and in hospitalized and non hospitalized patients.It means that invasive population was not differentiated from colonizing strains and indicated that any clone can be pathogenic in any infection site.This suggests also that endogenous flora of the host was the main source of infection and excludes the hypothesis that inter-patient transfer plays a role in the prevalence of at least most popular genotypes.
Moreover, some multilocus genotype G22 and G13 were more prevalent in invasive population.The genotype G22 were almost absent in out-population (vaginal and buccal samples).Our results suggested that some multilocus type may be stronger pathogen or that horizontal transfers can explain their highest prevalence inside the hospital.Furthermore, we notice that four patients (among 7 patients) infected with genotype G22 were hospitalized in 2009 in the same intensive care unit in the same period.This made that second hypothesis was more available and can be confirmed when analyzing a more large collection of isolates within and outside of the hospital.
It is also important to note that antifungal pressure exerted by some resistant genotypes can play a role in determining the population structure inside the hospital.It had been previously shown that there was no association between antifungal susceptibility and particular genotypes in previous study in Tunisia [3].Similarly, Lott et al. [11] affirmed that fluconazole resistance profiles from invasive and non invasive population were not significantly different and were not associated with a particular sequence type.In our study, we didn't found an emergence of any of the major multilocus type and fluconazole resistance in inside or in the outside population hospital except for isolates collected in 2009 (4 isolates) in which two of them showed a resistant phenotype (Figure 1).Eleven percent of C. Glabrata isolates in this study were fluconazole resistant.Among them, 80% were isolated from hospitalized patients (p ≥ 0.05).Selection pressure exerted in patients inside the hospital perhaps due to antifungal treatment may be responsible.
We can conclude that very little differentiation exists between both hospitalized and non hospitalized patients, invasive and colonizing population and the dissemination to multiple patients from a single source (health care worker, equipment) could be an important factor in the transmission of this yeast in our hospital.

Figure 2 :
Figure 2: Neighbor-joining dendrogram for the 62 genotypes identified in this study.Major groups previously identified are indicated.

Table 1 :
Number of isolates in subdivised population, MIC, Number of isolates with major multilocus type and Desequilibrium linkage analyses.

Genetic Structure of Candida glabrata Isolates from Hospitalized and Non-Hospitalized Patients in Sfax-Tunisia Copyright: © 2015 Abbes et al. pairs
of loci are compatible with strict clonality.

Table 2 :
Genotype distribution of multilocus type identified in our study.