Portrait of Matrix Gene Expression in Candida glabrata Biofilms with Stress Induced by Different Drugs

(1) Background: Candida glabrata is one of the most significant Candida species associated with severe cases of candidiasis. Biofilm formation is an important feature, closely associated with antifungal resistance, involving alterations of gene expression or mutations, which can result in the failure of antifungal treatments. Hence, the main goal of this work was to evaluate the role of a set of genes, associated with matrix production, in the resistance of C. glabrata biofilms to antifungal drugs. (2) Methods: the determination of the expression of BGL2, XOG1, FKS1, FKS2, GAS2, KNH1, UGP1, and MNN2 genes in 48-h biofilm’s cells of three C. glabrata strains was performed through quantitative real-time PCR (RT-qPCR), after contact with Fluconazole (Flu), Amphotericin B (AmB), Caspofungin (Csf), or Micafungin (Mcf). (3) Results: Mcf induced a general overexpression of the selected genes. It was verified that the genes related to the production of β-1,3-glucans (BGL2, XOG1, GAS2) had the highest expressions. (4) Conclusion: though β-1,6-glucans and mannans are an essential part of the cell and biofilm matrix, C. glabrata biofilm cells seem to contribute more to the replacement of β-1,3-glucans. Thus, these biopolymers seem to have a greater impact on the biofilm matrix composition and, consequently, a role in the biofilm resistance to antifungal drugs.


Introduction
Fungal infections continue to increase worldwide, particularly among immunosuppressed patients, individuals under prolonged hospitalization, catheterization, or continued antimicrobial treatments [1][2][3]. Candida spp. are the commonest fungal species involved in these diseases. Candida albicans is the most isolated species, but Candida glabrata and Candida parapsilosis are the second most isolated species in the United States of America and Europe, respectively [1,4,5]. Though C. glabrata does not have the capacity to form hyphae and pseudohyphae or to secret proteases, this species has other virulence factors, such as the ability to secrete phospholipases, lipases, and haemolysins and, importantly, the capacity to form biofilms [6][7][8]. These factors highly contribute to a high aggressiveness, resulting in a low therapeutic response and severe cases of recurrent candidiasis [8,9]. Biofilms are communities of microorganisms that colonize tissues and indwelling medical devices, embedded in an extracellular matrix [10,11]. These heterogeneous structures provide high resistance to antifungal therapy and strong host immune responses [7,8,12]. C. glabrata has shown to form a compact biofilm structure in different multilayers [6,7], with proteins, carbohydrates, and ergosterol into their matrices [6,7,13].
Various reports have shown the presence of β-1,3 glucans in the biofilm matrices of C. albicans [14][15][16][17]. Interestingly, it has been demonstrated that an increase in cell wall glucan was associated with biofilm growth [14] and, more recently, β-1,3 glucans were shown to be also present in the matrices of C. glabrata biofilms [13,18,19]. This specific carbohydrate has been associated with a general increase of extracellular

Organisms
Three strains of C. glabrata were used in the course of this study: One reference strain (C. glabrata ATCC 2001) from the American Type Culture Collection (Manassas, VA, USA), one strain recovered from the urinary tract (C. glabrata 562123) of a patient, and one strain recovered from the vaginal tract of a patient (C. glabrata 534784) in the Hospital Escala, Braga, Portugal. The identity of all isolates was confirmed using CHROMagar TM Candida (CHROMagar TM , Paris, France) and by PCR-based sequencing using specific primers (ITS1 and ITS4) against the 5.8 s subunit gene reference [60]. The PCR products were sequenced using the ABI-PRISM Big Dye terminator cycle sequencing kit (Perkin Elmer, Applied Biosystems, Warrington, UK).

Growth Conditions
For each experiment, C. glabrata ATCC2001, C. glabrata 534784, and C. glabrata 562123 strains were subcultured on Sabouraud dextrose agar (SDA) (Merck, Darmstadt, Germany) for 24 h at 37 • C. The cells were then inoculated in Sabouraud dextrose broth (SDB) (Merck) and incubated for 18 h at 37 • C under agitation at 120 rpm. After incubation, the cells were harvested by centrifugation at 3000× g (Thermo Scientific, CL10, Hampton, NH, USA) for 10 min at 4 • C and washed twice with phosphate buffered saline (PBS, pH = 7.5). The cell pellets were then suspended in Roswell Park memorial institute (RPMI), and the cellular density was adjusted to 1 × 10 5 cells/mL, using a Neubauer counting chamber.

Biofilm Formation
The minimum biofilm eradicatory concentration (MBEC) values were previously determined by the group, according to the European committee on antimicrobial susceptibility testing (EUCAST) guidelines [61,62]. For biofilm formation, standardized cell suspensions (1000 µL) were placed into selected wells of 24-wells polystyrene microtiter plates (Orange Scientific, Braine-l'Alleud, Belgium). At 24 h, 500 µL of RPMI-1640 was removed, and an equal volume of fresh RPMI-1640 plus the antifungal solution was added, on the basis of the MBEC values determined and indicated in bold in Table 1 (2× concentrated). The plates were incubated at 37 • C for additional 24 h at 120 rpm. RPMI-1640 containing only the antifungal agent was used as a negative control. As a positive control, cell suspensions were tested in the absence of the antifungal agent [18]. Genes related to the production of biofilm matrix components (β-1,3, β-1,6 glucans, and mannans)-BGL2, FKS1, FKS2, GAS2, KNH1, UGP1, XOG1 and MNN2-were selected for this study. The gene sequences of interest were obtained from Candida Genome Database [63] and the primers for quantitative real-time PCR (RT-qPCR) were designed using Primer 3 [64] web-based software and are listed in Table 2. ACT1 was chosen as a housekeeping gene. In order to verify the specificity of each primer pair for its corresponding target gene, the PCR products were first amplified from C. glabrata ATCC2001.

Preparation of Biofilm Cells for RNA Extraction
After biofilm formation, the medium was eliminated, and the wells were washed with sterile water to remove non-adherent cells. The biofilms were scraped from the wells with 1 mL of sterile water and sonicated (Ultrasonic Processor, Cole-Parmer, IL, USA) for 30 s at 30 W to separate the cells from the biofilm matrix. The cells were harvested by centrifugation at 8000× g for 5 min at 4 • C [18].

RNA Extraction
RNA extraction was performed using PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA, USA). Prior to RNA extraction, a lysis buffer from PureLink RNA Mini kit was prepared by adding 1% of ß-mercaptoethanol to the supplied buffer solution. Then, 500 µL of lysis buffer containing glass beads (0.5 mm diameter) was added to each pellet. The cell suspensions were homogenized twice for 30 s using a Mini-Bead-Beater-8 (Stratech Scientific, Soham, UK). After cell disruption, the PureLink RNA Mini Kit (Invitrogen) was used for total RNA extraction according to the manufacturer's recommended protocol. To avoid potential DNA contamination, the samples were treated with RNase-Free DNase I (Invitrogen) [18]. Table 2. Primers, targets used, and specific function of the genes used for the expression analysis.

Quantitative Real-Time PCR
RT-qPCR (CFX96 Real-Time PCR System, Bio-Rad) was performed to determine the relative levels of all genes mRNA transcripts in the RNA samples, with ACT1 used as a reference Candida housekeeping gene. Each reaction mixture consisted of a working concentration of SoFast EvaGreen Supermix (Bio-Rad), 50 µM forward and reverse primers, and 4 µL cDNA, in a final reaction volume of 20 µL. Negative controls (water) as well as non-transcriptase reverse controls (NRT) were included in each run [18]. The relative quantification of gene expression was performed by the 2 −∆C T method [65]. Each reaction was performed in triplicate, and mean values of relative expression were determined for each gene. The results are presented after calculation of 2 −∆C T .

Statistical Analysis
All experiments were repeated three times in independent assays. The results were compared using one-way analysis of variance (ANOVA), Dunnett's multiple comparisons tests, using GraphPad TM Prism TM 7 software (GraphPad Software, San Diego, CA, USA). All tests were performed with a confidence level of 95%. In order to determine the similarity of the strains' gene profiles, the Pearson Correlation Coefficient (r) was also applied.

Results and Discussion
Candidaemia related to C. glabrata has been increasing in the last years in parallel with its high drug resistance, particularly to the azole antifungal class [1,20,66]. Biofilms of C. glabrata are highly recalcitrant to treatments with antifungal agents as a consequence of multiple resistance mechanisms, such as those linked to the presence of a strong net of exopolysaccharydes and other biopolymers that protect the cells and hinder the diffusion of the drugs [1,15,[67][68][69]. In order to stress C. glabrata biofilm cells, four antifungals were applied (at concentrations based on MBECs values, Table 1) in pre-formed biofilms, and then an evaluation of biofilms' matrix gene expression was performed and compared with the expression of a housekeeping gene. Figure 1 shows the heatmap with the results of the RT-qPCR expression profiling of biofilm cells of C. glabrata ATCC2001 (A), C. glabrata 562123 (B), and C. glabrata 534784 (C) in the presence of antifungal drugs. The final data are presented in fold-change (FC) in comparison to the expression of the housekeeping gene (2 −∆C T ) [70]. Generally, BGL2, FKS1, FKS2, GAS2, and XOG1 displayed higher expression levels in biofilm cells in response to the drugs and, by contrast, KNH1, UPG1, and MNN2 displayed minor expression changes ( Figure 1 and Table 3).

Gene
Candida  In an important report, Taff et al. [19] concluded that mutants of C. albicans unable to produce Bgl2 and Xog1 enzymes did not show perturbations in the cell wall glucan composition of biofilm cells, and that these enzymes were not necessary for filamentation or biofilm formation. However, the biofilms had a reduced matrix glucan content, reduced total matrix biomass accumulation, and improved susceptibility to antifungal drug therapy [19]. Similarly, Li et al. [71] showed that, in C. albicans' persister cells (frequent in biofilms [72,73]), there was an increased expression of cell wall integrity proteins such as Xog1 and Bgl2. These studies recognized a biofilm-specific pathway involving Bgl2 and Xog1 (and Phr1) enzymes and affecting matrix delivery, by which these enzymes release and modify cell wall glucan for deposition in the extracellular space; however, an alternative explanation is that these enzymes act in the extracellular space, being crucial for mature matrix organization and function [19]. These enzymes have been localized in the cell wall, supporting the hypothesis of their activity in the cell wall, but have also secretion sequences that support an extracellular function. As seen earlier, BGL2 is one of the glucan modifying genes for glucan delivery, and XOG1 is a glucanase [19], necessary for modification and delivery of carbohydrates to the mature biofilm matrix. Without delivery and accumulation of matrix glucan, the biofilms exhibit enhanced susceptibility to antifungal drugs [19]. The change in the regulation of BGL2 and XOG1 in the biofilm cells of C. glabrata after drug treatment that we observed is interpreted as a response of the biofilm cells to the reduction of biofilm matrix, specifically of β-1,3-glucans, and it has been described before [7,13,18].
The GAS gene family is also a regulator in the production of β-1,3-glucan, and Gas2 is a glycosylphosphatidylinositol (GPI)-anchored cell surface protein [31] involved in the production of β-1,3-glucan in C. glabrata [29,30]. Gas2 is a documented putative carbohydrate-active enzyme and consequently it can alter the cell wall polysaccharides in order to build and remodel the cell wall glycan network during growth in C. glabrata [29]. In C. glabrata ATCC2001, GAS2 was highly expressed in the non-treated group and after Csf contact (FC: 5.72), while Flu, AmB, and Mcf led to its downregulation. The clinical isolates upregulated the gene in all conditions, except for AmB treatment of C. glabrata 562123 ( Figure 1 and Table 3, all p < 0.0001). Hence, when analyzing the results of C. glabrata 562123 and C. glabrata 534784, the GAS2 network seems to be activated also after glycan's loss following drug treatment, in order to replace the lack of 1,3-β-glucans and re-establish biofilm cells' homeostasis. All results compared with those from untreated cells (controls) were statistically significant (p < 0.0001).
The resistance to echinocandins increased from 4.9% to 12.3% between 2001 and 2010 [74] with a rapid development of FKS mutations in Candida spp., especially in C. glabrata [75,76]. The amino acid substitutions occurring in FKS1 [30,[77][78][79] and FKS2 [30,80] are directly related to the resistance to this class of drugs: acquired FKS mutations [81] are reported to confer low β-(1,3)-D-glucan synthase sensitivity and to increase the minimum inhibitory concentration (MIC) values, which are related to clinical failure [82]; intrinsic FKS mutations, also lead to elevated MIC levels but have a weaker effect on the reduction of β-(1,3)-D-glucan synthase sensitivity [82][83][84][85]. Generally, in the C. glabrata strains, a 24 h contact with both echinocandins upregulated FKS1 and FKS2 genes and, in the reference strain, the presence of Mcf upregulated FKS1. More specifically, the results showed that all strains upregulated the expression of FKS1 after drug exposure (statistically significant), with the exception of Csf in C. glabrata ATCC2001 (FC: 0.07 p < 0.001) and Csf in C. glabrata 562123 (FC: 0.08; p < 0.0001). For FKS2, its overexpression was observed for almost all treatments in the three strains, excluding following AmB treatment in C. glabrata ATCC2001 (FC: 0.28; p < 0.001) and C. glabrata 562123 (FC: 0.55; p < 0.05). C. glabrata 534784 revealed, again, to have the highest capacity to overexpress both genes in response to drug stress. These differences among the strains may be related to the described Candida spp. intra-strains variations [62]. Bizerra et al. [76] reported the occurrence of a mutation associated with the resistance phenotype against echinocandins in C. glabrata isolated from a single cancer patient with candidemia exposed to antifungal prophylaxis with Mcf. Arendrup et al. [86] revealed that Mcf MICs of C. glabrata FKS hot spot mutant isolates were less raised than those obtained for the other echinocandins, showing that the efficacy of Mcf could be differentially dependent on specific FKS genes mutations. These reports mention singularities regarding the FKS gene and Mcf, which can also be observed in our results (Figure 1 and Table 3). Interestingly, up and downregulations of FKS1 and FKS2 were similar in the clinical isolates and parallel to those observed for BGL2, which makes sense, since this gene has shown to perform, with XOG1 (and PHR1), in a complementary manner in order to distribute the matrix downstream of the primary β-1,3 glucan synthase encoded by FKS1 [19]. Previous investigations also found elevated transcript levels of FKS1, BGL2, and XOG1 during in vivo C. albicans biofilm growth when compared to planktonic growth, which is consistent with our results and with a role in a biofilm-specific function, such as matrix formation [87,88].
The overexpressed values obtained for BGL1, XOG1, FKS1, FKS2, and GAS2 after the stress conditions induced by most antifungals endorse the impact of β-1,3-glucans in the maintenance of the cell and biofilm matrix structure.
Sequencing studies have shown that C. glabrata is more closely related to S. cerevisiae than to C. albicans [89], with some genes functionally interchangeable among the two species [90,91]. An important component of the cell wall and the biofilm matrix is β-1,6-glucan, which is regulated by several genes, such as KNH1. Preceding studies have demonstrated that the KNH1 homologs are essential components of β-1,6-glucan synthesis in C. glabrata [35,36,38]. In S. cerevisiae, many genes involved in β-1,6-glucan synthesis were isolated through mutations (kre [killer resistant] mutations) that are responsible for the resistance to the K1 killer toxin, which kills sensitive yeast cells after binding to β-1,6-glucan [35,38,92]. Dijkgraaf and colleagues [35] reported that the disruption of both KRE9 and KNH1 was synthetically lethal for C. glabrata, demonstrating the importance of these genes in the maintenance of cell structure. In the present study, after a drug stress, all C. glabrata strains upregulated this gene (Figure 1 and Table 3), indicating an effort to replace these β-1,6-glucans after losses due to the aggression of the antifungals, confirming also a certain degree of relevance of these elements in the cell wall and biofilm matrix of C. glabrata [35,36,38]. C. glabrata ATCC2001 showed to upregulate the KNH1 gene in the presence of antifungal drugs, and C. glabrata 562123 indicated an identical pattern by marginally increasing KNH1 gene expression in these conditions. Compared to the other two strains, with the exception of Flu treatment (FC: 2.08; p < 0.0001), KNH1 showed a different regulation in the vaginal tract strain (Figure 1 and Table 3). C. glabrata 534784 demonstrated to have the highest upregulation capacity, presenting overexpression almost for all genes (Figure 1 and Table 3).
During glucose starvation, a set of genes orthologous to S. cerevisiae is induced in C. glabrata, including UGP1, related to the β-1,6-D-glucan biosynthetic process [39,40], which shows that the environmental stress response is conserved between S. cerevisiae and C. glabrata [39]. UPG1 showed to have the lowest expression, compared with other genes and controls. Nonetheless, except for one condition and strain, in the presence of antifungal drugs, several overexpression states were observed ( Figure 1 and Table 3). The reference strain displayed overexpression in all conditions, with the highest gene upregulation occurring in the presence of AmB (FC: 0.17; p < 0.0001) and the lowest in the presence of Csf (FC: 0.01; non-significant); the urinary tract strain also revealed limited gene upregulation in the presence of Csf and AmB and the highest expression in the presence of Mcf (0.15; p < 0.0001). Generally, C. glabrata 534784 demonstrated the highest FC expression in all conditions. The lowest gene upregulation was observed in biofilm cells stressed by Csf (FC: 0.33; p < 0.0001). Srikantha and colleagues [91] identified a set of genes that are upregulated by the transcription factor Bcr1, involved in impermeability, impenetrability, and drug resistance of C. albicans' biofilms. The authors concluded that the induction of Bcr1 overexpression in weak biofilms of C. albicans conferred those three characteristics and, in these cases, UGP1 gene was downregulated [91]. This result supports the FC expression we obtained: since C. glabrata biofilms were weakened by the drugs, UGP1 expression was increased in order to balance this defect (as seen with KNH1). The overexpression values we obtained for both KNH1 and UPG1 point to the relevance of β-1,6-glucans in the maintenance of a good cell and matrix structure.
Regarding mannans regulation, all strains showed a low or moderate expression of MNN2 in the controls (non-treated cells), but relevant expression changes arose in the presence of all drugs, particularly when Mcf was added ( Figure 1 and Table 3). C. glabrata ATCC2001 demonstrated the lowest expression in the control group, among all strains ( Figure 1 and Table 3). The urinary strain presented the lowest gene expression, when compared to the untreated group ( Figure 1 and Table 3). Flu and Mcf induced the highest MNN2 values (FC: 0.34 and 1.21, respectively, both p < 0.0001), while AmB and Csf were associated with the lowest expressions (FC: 0.24, p < 0.0001 and 0.13, p < 0.0005). For C. glabrata ATCC2001 and C. glabrata 534784, the weakest effects were associated with the biofilm cells that were stressed by AmB (FC: 0.13 and 0.71 respectively, both p < 0.0001). When Mcf was applied, the biofilm cells of the vaginal strain showed a strong response to the stress, compared to the other two strains (FC: 7.03; p < 0.0001). Our team has also found that C. glabrata ATCC2001 increased the amounts of mannans on its cell walls in the presence of these drugs (data not shown), revealing a possible adaptation of the cells to the stress caused by the antifungal drugs. Other studies reported analogous adjustments of the cell walls after environmental drug stress, which has been related to high antifungal resistance events [1,2,19,[92][93][94], supporting these results.
Interestingly, and when compared to the rest of the genes, the present results demonstrate that KNH1, UGP1, and MNN2 had the lowest values of FC expression. This seems to indicate that, although β-1,6-glucans are an important part of the cell and biofilm matrix, the cells appear to invest more in replacing the lost β-1,3-glucans, leading to consider that these components have a greater significance in the maintenance of the homeostasis of the biofilm matrix and the biofilm cells. In fact, in studies developed in our group [13,18], the total polysaccharides and β-1,3-glucans concentrations increased significantly in C. glabrata biofilm matrices after Flu, AmB, and Mcf contact. These higher concentrations in β-1,3 glucans content might explain part of the main biofilm resistance to the drugs that was formerly described [7,[95][96][97].
Finally, downregulation of most genes and strains happened in the presence of AmB whereas, in opposition, Mcf induced the main overexpression alterations (Figure 1 and Table 3). AmB is a fungicidal drug and the most important antifungal polyene used for the treatment of systemic fungal infections [98,99]. This drug binds to the ergosterol of the cell membrane but also induces oxidative stress. This explains the existence of a still low reported rate of resistance and the good effectiveness of AmB [1,13,[100][101][102][103][104]. Also, this low resistance may be associated with the lower gene expression effects that we detected after AmB exposure in C. glabrata ( Figure 1 and Table 3). In opposition, the most acute upregulations occurred in the presence of echinocandins and, particularly, when Mcf was applied. This class of antifungals act by inhibiting β-1,3-glucan synthesis [1,100,105], which affects cell wall and matrix composition. By overexpressing the genes related to β-1,3-glucan synthesis (BGL2, FKS1, FKS2, GAS2, XOG1), the cells were attempting to compensate and replace the β-1,3-glucan losses in their matrices induced by the drugs and, thus, protect and decrease their susceptibility to the antifungals [19]. This general increase in total carbohydrates and specifically in β-1,3-glucans in Candida spp. biofilm matrices has already been described [7,13,18].
Regarding the correlation between the gene expression profiles in C. glabrata, the results based on the r are displayed in Table 4. The results showed a strong positive correlation (r near 1) between the response profiles of BGL2, XOG1, FKS1, and MNN2 gene expression in the three strains, which means that up and downregulation had a high tendency to occur similarly in all strains. The scores of the r for the profile of the FKS2 gene revealed a moderate positive correlation between the reference strain (C. glabrata ATCC2001) and the isolates (C. glabrata 562123 and C. glabrata 534784). This indicates that, although the correlation was positive, it was weak, and the profiles of the gene response were variable in the three strains. On the other hand, the clinical isolates showed strong positive correlation between the expression profiles of this gene. C. glabrata ATCC2001 demonstrated a moderate positive correlation between the expression profiles of the GAS2 gene. C. glabrata 562123 and C. glabrata 534784 had a strong positive correlation between the expression profiles of the GAS2 gene. KNH1 gene was the most variable and difficult gene to correlate between the strains. The reference and the 562123 strain showed a moderate correlation, whereas the reference and the 534784 strain showed a strong correlation, and the clinical isolates showed the only weak correlation detected in this study. As for the UPG1 gene, although there was a negative correlation, the association between its expression in ATCC2001 and in the clinical isolates can be considered weak. Between the isolates, it was determined that UPG1 up and downregulation had a high tendency to occur similarly in all strains (thus showing strong correlation). In summary, BGL2, XOG1, FKS1, and MNN2 appeared to be the genes presenting the most similar responses to antifungal drugs within the transcriptome of the three strains; also, the clinical isolates appeared to be nearer each other than to the reference strain. Once more, β-1,3-glucan synthesis was identified as important in C. glabrata (three of the four genes affected are responsible for β-1,3-glucan production). These similarities among the two clinical strains may be due to the fact that both were derived from a hospital environment, and it is probable that they had already been challenged by several drugs, so their responses were prompter compared to the reference strain that is a wild type strain.

Conclusions
The in vitro high-dose paradox associated with Candida spp. isolates is being increasingly reported and connected to slightly elevated MICs, potentially contributing to clinical resistance and failure of antifungal treatments. These drug tolerance and adaptive mechanisms are highly related to Candida spp. biofilm forms. C. glabrata extracellular matrix is crucial for mature biofilm formation, not only contributing to the adhesive nature of the biofilm cells, but also protecting the cells from antifungal agents and from the host immune system. Understanding the production of the biofilm matrix components and the associated delivery processes is important for the development of effective biofilm therapies. All stakeholders in this process represent potentially attractive targets for detection of and therapeutic interventions against candidiasis.