Multidrug-resistant Candida glabrata strains obtained by induction of anidulafungin resistance in planktonic and biofilm cells

1Laboratório de Micologia Aplicada, Faculdade de Farmácia, Universidade Federal do Rio Grande do Sul, UFRGS, Porto Alegre, Rio Grande do Sul, Brazil, 2Programa de Pós-Graduação em Microbiologia Agrícola e do Ambiente, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil, 3Departamento de Química Orgânica, Instituto de Química, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil


INTRODUCTION
In recent decades, the frequency of invasive yeast infections has significantly increased with the increase in the number of patients undergoing immunosuppressive therapy, the use of broad-spectrum antimicrobials, and the frequency of organ transplants (Silva et al., 2012). Systemic infections caused by Candida glabrata are characterized by a high mortality rate and are difficult to treat due to reduced susceptibility to azole antifungals, particularly fluconazole (McCornack, Perry, 2005;Katiyar et al., 2012;Rodrigues, Silva, Henriques, 2014). This yeast can colonize host tissues and abiotic surfaces (catheters and prostheses) where it develops multi-layer growth characterized as a biofilm (Katiyar et al., 2012).
Biofilms are communities of microorganisms adhering to a surface and embedded in a polymeric extracellular matrix (Donlan, Costerton, 2002) and considered the most prevalent form of growth in the environment they inhabit. Biofilm formation is an essential virulence factor for Candida species, and it confers significant resistance to antifungal therapy by limiting the penetration of substances into the matrix and protecting the cells against the immune response of the host (Mukherjee, Chandra J, 2004).
Considering the importance of invasive mycoses and their current epidemiology, it is necessary to develop multiple therapeutic strategies. A few years ago, therapeutic management was performed with azoles and amphotericin B. However, considering the increasing azole resistance, mainly caused by non-albicans Candida species and due to renal toxicity of amphotericin B, a new pharmacological group, the echinocandins, emerged as a novel therapeutic option (Pham et al., 2014).
Echinocandins inhibit the 1,3-β-D-glucan synthase (GS) complex that catalyzes the biosynthesis of 1,3-β-Dglucan, the primary glucan component of the fungal cell wall. GS, an enzyme, is complex and has at least two subunits: Fksp and Rho1p. Among these, Fksp, encoded by three paralogous genes (FKS1, FKS2, and FKS3), is the catalytic subunit and target of echinocandins. Resistance to these antifungal agents, resulting in therapeutic failure, has been linked to mutations in the Fksp subunit of GS (Riera et al., 2012).
Recently, the yeast C. glabrata, usually commensal, has emerged as a common cause of serious life-threatening fungal infections largely owing to their low susceptibility to azole antifungals widely used in the treatment of these diseases (Silva et al., 2012). Consequently, recent guidance indicates the use of echinocandins as the first-choice drug for the treatment of systemic infections of C. glabrata (Silva et al., 2012;Pham et al., 2014). However, studies indicate that C. glabrata resistance to echinocandins is increasing (Pham et al., 2014).
Fluconazole-resistant isolates of C. glabrata have an increased possibility of being non-susceptible to echinocandins and vice versa. C. glabrata can present point mutations and changes in the chromosome structure, possibly related to mechanisms of adaptation to environmental changes. These genomic modifications may be a result of a defense mechanism making this fungus rapidly resistant to multiple drugs after limited exposure. For C. glabrata non-sensible to azoles and echinocandins, the only antifungal agent available is amphotericin B. Consequently, a two-fold problem arises because infections increase with aging, and amphotericin B is not well tolerated by elderly patients (Spreghini et al., 2012).
Herein, we present a study of the induction of anidulafungin resistance in planktonic and sessile cells of C. glabrata and investigate the development of fluconazole cross-resistance.

Biofilm formation assay on tracheal catheter material
The applied methodology is an adaptation of a procedure reported in the literature Tondo et al., 2010). Initially, the strains were cultured on Sabouraud agar for 24 h at 35 ºC. Seven colonies were added to 2 mL TSB followed by incubation for 24 h at 35 ºC. Further, 1-× 1-cm 2 polyvinyl polychloride (PVC, Mark Med, number 12) catheter fragments were added to 9 mL peptone water containing 1 mL standardized inoculum in TSB followed by incubation for 96 h at 35 °C. Then, the catheter fragments were washed thrice with peptone water to remove poorly adhered cells. The fragments were added to other flasks containing 50 mL peptone water, and the adhered cells were released by sonication at 40 KHz for 10 min, the resulting water following sonication was diluted (dilution factor, 10), allowing the determination of colony forming units (CFU), and 20 μL of each dilution was plated on Sabouraud agar. Next, the plates were incubated for 24 h at 35 °C, and the values of CFU/cm 2 were determined. All counts were performed in triplicate. The best biofilm-forming strain cells obtained by this test were cultivated on Sabouraud agar and stored for further use.

Planktonic and sessile minimum inhibitory concentration (MIC)
MICs of the planktonic and sessile cells of C. glabrata were determined in microplates by a broth microdilution method, as proposed by Clinical and Laboratory Standards Institute (CLSI), in accordance with protocol M27-A3 (CLSI, 2008), following the updates of document M27-S4 (CLSI, 2012). The antifungal was tested at concentrations of 0.0075−4.0 μg/mL.

Induction of anidulafungin resistance in planktonic and sessile cells and evaluation of its stability
This test was performed by exposing the strains to successively higher concentrations of antifungal, as described by Fekete-forgacs, Gyure, and Lenkey (2000), with slight modifications. One colony of each C. glabrata strain, anidulafungin-sensitive [≤ 0.12 µg/mL for anidulafungin was considered as susceptible and ≥0.5 µg/mL as resistant, M27-S4 (CLSI, 2012)], was incubated with 10 mL RPMI 1640 containing 4% glucose for 24 h at 35 °C. After centrifugation, supernatant was discarded, leaving only the yeast cells in the vials. Further, RPMI 1640 was added until a final absorbance of 0.1 (λ = 640 nm) was obtained. After 10-h incubation, the material was centrifuged, and supernatant was discarded; 10 mL anidulafungin was diluted with RPMI to the desired concentration (MIC/8). After 14-h incubation, the previous procedure was repeated three consecutive times but with a 24-h incubation interval. After the third incubation, 20 μL of each culture was grown on Sabouraud agar for 24 h. A sample of cells was subcultured for MIC determination and stored for further testing. After another round of centrifugation and discarding of supernatant, the antifungal was added at the MIC/8 concentration until the final absorbance of 0.1 was obtained. After 10-h incubation, the culture was centrifuged again, and supernatant was discarded. Then, 10 mL MIC/4 concentration of anidulafungin was added and incubated for 14 h, as described above. The experiment was repeated doubling the concentration of anidulafungin until the final concentration of 0.25 μg/mL was obtained for all strains.

Fluconazole cross-resistance detection
Fluconazole cross-resistance was assessed by MIC determination of anidulafungin-resistant strains and compared with MIC prior to the induction of in vitro resistance.

RESULTS
The biofilm formation assay demonstrated that all C. glabrata strains could form a biofilm on the tracheal catheter material (Table I). This screening for the identification of optimum biofilm formation indicated seven outstanding biofilm-forming strains (RL12, CG18S, CG40039, RL09m, RL22, RL25, and RL26), which were selected for the experiments of induction of anidulafungin resistance. These strains were selected based on the results that showed the highest colony count (CFU/cm 2 and log).
MICs of the planktonic and sessile cells of C. glabrata determined by a broth microdilution method, as proposed by the CLSI, are presented in Table II. In general, the sessile cells presented slightly reduced susceptibility compared with the planktonic cells. MICs were in the range of <0.0075−0.03 μg/mL for the planktonic cells and 0.015−0.06 μg/mL for the sessile cells.
Following the induction of anidulafungin resistance against the sessile and planktonic forms of the C. glabrata strains, MICs were determined again. It was observed that six out of seven strains tested (86%) acquired strong anidulafungin resistance (Table II). Differences between MICs of the biofilm and planktonic cells after the induction of in vitro resistance were not noted, which was previously verified. In addition, morphological changes (phenotypic switching) to non-susceptible C. glabrata strains were observed. Furthermore, after 30-day subculturing of the resistant strains in anidulafungin-free medium, MICs indicated that resistance was maintained for six strains.
Moreover, the occurrence of cross-resistance to fluconazole was investigated by an MIC assay for anidulafungin-resistant C. glabrata strains. MICs were compared with anidulafungin-sensitive strains. Surprisingly, all strains with previous exposure to anidulafungin showed resistance to fluconazole (MIC ≥ 64 µg/mL; Table III). For the planktonic and sessile cells, very similar results were obtained.

DISCUSSION
According to protocol M27-A3 4 and considering the updates of document M27-S4 (CLSI, 2012), anidulafungin-susceptible isolates present MICs ≤ 0.12 μg/mL. Notably, all selected strains were sensitive to the antifungal tested, in both their planktonic and sessile forms (Table II). In previous studies described by Pfaller and collaborators,only 3.8% of the isolates of C. glabrata obtained from Asia were resistant to anidulafungin, whereas none of the isolates from Latin America was non-susceptible to this antifungal (Pfaller et al., 2012(Pfaller et al., , 2013. The sessile form of the strains was less or equally sensitive to anidulafungin when compared with the planktonic form (Table II), which corroborates the findings of Kucharikova et al. (2011), suggesting that C. glabrata biofilms formed in vitro or by animal models are resistant to azoles but are sensitive to echinocandins.
The induction of resistance resulted in strains strongly resistant to anidulafungin with MICs in the range of 1−2 μg/mL (Table II)  after in vitro exposure to higher concentrations of micafungin and anidulafungin. In their experiment, MICs of Sabouraud agar plates containing these antifungals were 0.06−4 μg/mL (micafungin) and 1−4 μg/mL (anidulafungin) (Bordallo-Cardona et al., 2017).

Sessile cells control
Biofilm cells have antimicrobial resistance characteristics that confer tolerance to drug concentrations 1000 times higher than the MIC determined for planktonic cells (D'enfert, Janbon, 2016). Thus, as a final result of the induction of in vitro resistance, larger MICs were expected for the sessile than for the planktonic cells. Surprisingly, this behavior was not observed in the present experiments (Table II). It was verified that biofilm and planktonic cells present the same MIC, not maintaining their susceptibility observed prior to the induction of in vitro resistance.
To analyze the stability of the developed resistance, the strains were harvested for 30 days with antifungalfree Sabouraud agar. The results showed the preservation of resistance for all strains. Previously, Borst and coworkers (2005) also demonstrated the maintenance of acquired resistance to C. glabrata for at least 4 months after the removal of fluconazole. This result is clinically relevant and worrying because, once resistant, these cells will continue to express resistance genes for an extended period. Eventually, antifungal treatment becomes ineffective, making the introduction of a second therapeutic drug necessary. In addition, there is a risk of the dissemination of resistance of these isolates to other patients, further limiting therapeutic options.
In addition, morphological changes (phenotypic switching) observed in the C. glabrata strains can be related to resistance. Phenotypic switching enables microorganisms to undergo rapid microevolution and to adapt to a constantly changing microenvironment. In addition, it facilitates pathoadaptation into the host (Jain, Hasan, Fries, 2008). Similar to C. albicans, C. glabrata is capable of expressing different cellular phenotypes in colonizing populations, including phenotype irregular wrinkle (Lachke et al., 2002). To C. albicans, this effect was correlated to high-level antifungal resistance, indicating that morphological change may have an effect on drug sensitivity (Vargas et al., 2000).
According to the CLSI, the C. glabrata strains having MICs ≥ 64 μg/mL are considered resistant to fluconazole; meanwhile, C. glabrata presenting MICs ≤ 32 μg/mL are classified as sensitive in a dose-dependent manner. It was observed that prior to the induction of in vitro resistance, the planktonic and sessile forms of the strains were all sensitive to fluconazole in a dosedependent manner with MICs in the range of 1−32 μg/mL (Table III). Interestingly, after the induction of anidulafungin resistance, this dose-dependent sensitivity to fluconazole was not maintained, making these strains resistant to this antifungal.
Fluconazole resistance has been associated with non-susceptibility to echinocandins in other studies (Pfaller et al., 2012(Pfaller et al., , 2013. Azole resistance often results from the overexpression of efflux pumps, which are not related to echinocandin resistance. However, C. glabrata has a unique ability to sequentially acquire and express mutations responsible for conferring resistance. Therefore, it is possible that previous treatments with fluconazole exert selective pressure, resulting in changes in the genome of the pathogen even in the absence of evident fluconazole resistance. These microorganisms may persist as colonizers in patients and emerge as resistant to echinocandins after subsequent exposure to antifungal agents (Vallabhaneni et al., 2015). In our study, the opposite was observed: the emergence of fluconazole-resistant strains after previous exposure to anidulafungin. This observation was also previously reported (Alexander et al., 2013), suggesting that previous therapy with azoles or echinocandins is predictive of resistance to both classes of antifungals.
Finally, most of the sessile cells had slightly reduced susceptibility when compared with the planktonic cells. The strains subjected to the induction test to anidulafungin resistance also developed strong anidulafungin and fluconazole resistances, corroborating reports of crossresistance between the classes of azoles and echinocandins. Clinically, this cross-resistance could lead to therapeutic failure in fluconazole-treated patients who were previously exposed to subinhibitory concentrations of anidulafungin for extended periods.

CONFLICT OF INTEREST
The authors declare that they have no conflict of interest.