**Susceptibility of the** *Candida haemulonii* **Complex to Echinocandins: Focus on Both Planktonic and Biofilm Life Styles and a Literature Review**

#### **Lívia S. Ramos 1, Laura N. Silva <sup>1</sup> , Marta H. Branquinha <sup>1</sup> and André L. S. Santos 1,2,\***


Received: 20 August 2020; Accepted: 30 September 2020; Published: 1 October 2020

**Abstract:** *Candida haemulonii* complex (*C. haemulonii*, *C. duobushaemulonii* and *C. haemulonii* var. *vulnera*) is well-known for its resistance profile to different available antifungal drugs. Although echinocandins are the most effective class of antifungal compounds against the *C. haemulonii* species complex, clinical isolates resistant to caspofungin, micafungin and anidulafungin have already been reported. In this work, we present a literature review regarding the effects of echinocandins on this emergent fungal complex. Published data has revealed that micafungin and anidulafungin were more effective than caspofungin against the species forming the *C. haemulonii* complex. Subsequently, we investigated the susceptibilities of both planktonic and biofilm forms of 12 Brazilian clinical isolates of the *C. haemulonii* complex towards caspofungin and micafungin (anidulafungin was unavailable). The planktonic cells of all the fungal isolates were susceptible to both of the test echinocandins. Interestingly, echinocandins caused a significant reduction in the biofilm metabolic activity (viability) of almost all fungal isolates (11/12, 91.7%). Generally, the biofilm biomasses were also affected (reduction range 20–60%) upon exposure to caspofungin and micafungin. This is the first report of the anti-biofilm action of echinocandins against the multidrug-resistant opportunistic pathogens comprising the *C. haemulonii* complex, and unveils the therapeutic potential of these compounds.

**Keywords:** *Candida haemulonii* complex; planktonic growth; biofilm formation; echinocandins; caspofungin; micafungin

#### **1. Introduction**

The members of the *Candida haemulonii* species complex (*C. haemulonii*, *C. duobushaemulonii* and *C. haemulonii* var. *vulnera*) are well-known for their (multi)drug-resistance towards several antifungal agents available in clinical practice. Resistance of the *C. haemulonii* complex to azoles (e.g., fluconazole, itraconazole and voriconazole) and polyenes (e.g., amphotericin B) has been documented extensively [1–7]. On the other hand, susceptibility to prescribed echinocandins (anidulafungin, caspofungin and micafungin) is commonly observed [7–11], although there have been some reports of clinical isolates being resistant to these compounds [5,12].

Echinocandins are the newest class of antifungal agents to be used in clinical practice, exhibiting fungicidal activity against yeasts as well as having a good safety profile [8]. In this sense, the guidelines of the Centers for Disease Control and Prevention (CDC, USA) strongly recommend that echinocandins should be the first choice for the treatment of candidemia in both neutropenic and non-neutropenic patients [9]. The mechanism of action of the echinocandins involves the noncompetitive inhibition of the enzyme β-(1,3)-d-glucan synthase, which is involved in the synthesis of the polysaccharide glucan, resulting in the loss of cell wall integrity and severe stress in the fungal wall [8].

The three clinically available echinocandins usually exhibit both in vitro and in vivo fungicidal activity against a variety of *Candida* species, including those that are intrinsically resistant to azoles or amphotericin B (e.g., *C. krusei*, *C. glabrata* and *C. lusitaniae*), and also emerging species (e.g., *C. famata* and *C. rugosa*) [10]. Additionally, the antifungal activity of echinocandins against *Candida* biofilms represents an aspect that should be highlighted, since microbial biofilm is considered a resistance structure that precludes efficient antimicrobial treatment [10]. For instance, both caspofungin and micafungin, at concentrations attainable in clinical treatments, were able to kill fungal cells in preformed biofilms of either *C. albicans* or *C. parapsilosis* [11]. Therapeutic concentrations of caspofungin and micafungin were active against the biofilms formed by isolates of *C. albicans* and *C. glabrata* recovered from cases of bloodstream infections, but not against *C. tropicalis*, demonstrating that species-specific differences can influence the outcome [12]. Corroborating these findings, caspofungin was also shown to be effective in the treatment and prevention of *C. albicans* biofilms in an in vivo murine model of central venous catheter-associated candidiasis [13].

Considering the aforementioned aspects, the aim of the present study was to evaluate the antifungal susceptibility of both planktonic- and biofilm-forming cells from 12 Brazilian clinical isolates comprising the *C. haemulonii* complex towards caspofungin and micafungin. Furthermore, we have performed a literature review concerning the susceptibility of the *C. haemulonii* species complex towards echinocandins in order to present a comprehensive summary of this field.

#### **2. Materials and Methods**

#### *2.1. Microorganisms and Growth Conditions*

Twelve clinical fungal isolates, previously identified by molecular methods [6], belonging to the *C. haemulonii* species complex were used in the present study: five isolates of *C. haemulonii* (LIP*Ch*2 recovered from the sole of the foot, GenBank accession number KJ476194; LIP*Ch*3 from a toe nail, KJ476195; LIP*Ch*4 from a finger nail, KJ476196; LIP*Ch*7 from a toe nail, KJ476199; LIP*Ch*12 from blood, KJ476204), four isolates of *C. duobushaemulonii* (LIP*Ch*1 from finger nail, KJ476193; LIP*Ch*6 from a toe nail, KJ476198; LIP*Ch*8 from blood, KJ476200 and LIP*Ch*10 from bronchoalveolar lavage, KJ476202) and three isolates of *C. haemulonii* var. *vulnera* (LIP*Ch*5 from a toe nail, KJ476197; LIP*Ch*9 from urine, KJ476201 and LIP*Ch*11 from blood, KJ476203) [6]. In all experiments, Sabouraud dextrose medium was used to culture the fungal isolates at 37 ◦C for 48 h under constant agitation (200 rpm). Yeasts were counted in a Neubauer chamber.

#### *2.2. Determination of Minimal Inhibitory Concentration (MIC)*

Antifungal susceptibility testing, using the planktonic cells of *C. haemulonii* species complex, against caspofungin and micafungin (Sigma-Aldrich, St. Louis, MO, USA) was performed according to the broth microdilution technique standardized in the M27-Ed4 protocol [14] and interpreted according to the M27-S3 document published by the Clinical and Laboratory Standards Institute (CLSI) [15]. *C. krusei* (ATCC 6258) and *C. parapsilosis* (ATCC 22019) were used as quality control isolates in each test as directed by the CLSI. The clinical breakpoints to echinocandins are detailed below.

#### *2.3. Echinocandins' Breakpoints*

Until now, there have been no established breakpoints for echinocandins (or any other antifungal class) regarding the species belonging to the *C. haemulonii* complex. To overcome this problem, researchers working with this fungal complex, as well as "newly identified" *Candida* species, have generally been using a comparative perspective in order to interpret and discuss antifungal

susceptibilities. Results are normally presented as CLSI breakpoints which have been established for the *Candida* genus (CLSI document M27S3 [15]) in order to have a minimum (even if not precise) parameter to interpret this kind of experiment. Alternatively, a possible option is to compare the MIC values of *C. haemulonii* complex with the breakpoints established for non-*albicans Candida* species (e.g., *C. glabrata*, *C. tropicalis*, *C. krusei*, *C. parapsilosis* and *C. guilliermondii*) as recently suggested by the CLSI (document M27S4 [16] and protocol M60 [17]). However, this approach varies depending on the particular *Candida* species, since each presents its own breakpoint for each of the echinocandin drugs used. Moreover, the CDC (USA) recently published on its website (https://www.cdc.gov/fungal/candida-auris/c-auris-antifungal.html) a proposal of echinocandins' breakpoints for *C. auris*, a phylogenetically related species to the *C. haemulonii* complex, as follows: resistant breakpoint for caspofungin is ≥2 mg/L and for micafungin and anidulafungin, ≥4 mg/L. After contemplating these various viewpoints, we chose to use, herein, the breakpoints available for *Candida* spp. in the CLSI document M27-S3 [15], which considers as susceptible the strains having MIC values ≤2 mg/L and non-susceptible those with MIC values >2 mg/L for the three clinically available echinocandins; a MIC summary table was prepared.

#### *2.4. E*ff*ects of Echinocandins on the Biofilm Formed by the C. haemulonii Species Complex*

Fungal suspensions in Sabouraud broth (200 μL containing 106 yeast cells) were transferred into each well of a flat-bottom 96-well polystyrene microtiter plate and incubated without agitation at 37 ◦C for 48 h, which has been shown to be the best incubation time for biofilm formation by species belonging to the *C. haemulonii* complex [18]. Afterwards, the biofilm supernatant fluids were carefully removed, washed once with sterile phosphate-buffered saline (PBS; 10 mM NaH2PO4, 10 mM Na2HPO4, 150 mM NaCl, pH 7.2) and then 200 μL of Roswell Park Memorial Institute Medium (RPMI) 1640 medium containing different concentrations of echinocandins (range 0.25–8 mg/L) were added to each well. RPMI 1640 medium without echinocandins was used as a positive control and medium-only blanks were used as the negative control. The biofilms were then incubated at 37 ◦C for an additional 48 h. Afterwards, the supernatant fluids were carefully removed and the wells were washed twice with PBS to remove any non-adherent cells. Finally, two classic biofilm parameters (biomass and metabolic activity/viability) were measured as described below. The results were expressed as percentage of reduction of both viability and biomass. The minimal biofilm eradication concentration (MBEC) was achieved, considering the lowest concentration of each echinocandin capable of causing a 50% reduction in the biofilm viability [19].

#### 2.4.1. Viability Assay

The viability of the fungal cells forming the biofilm was determined using a colorimetric assay that measures the metabolic reduction of 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide (XTT; Sigma-Aldrich) to a water-soluble brown formazan product [20,21]. A XTT/menadione solution was prepared as follows: 2 mg of XTT was dissolved in 10 mL of pre-warmed PBS solution supplemented with 100 μL of a menadione stock solution (made by dissolving 55 mg of menadione in 100 mL of acetone). The XTT/menadione solution (200 μL) was added to all wells containing the biofilms (see Section 2.4 above) and incubated in the dark at 37 ◦C for 3 h. One hundred microliters of the supernatant from each well were then transferred to a new microplate and the colorimetric readings were measured at 492 nm using a microplate reader (SpectraMax M3; Molecular Devices, Sunnyvale, CA, USA) [21].

#### 2.4.2. Biomass Measurement

Biomass quantification was assessed as described by Peeters et al. [20]. Firstly, biofilms (see Section 2.4 above) were fixed by adding 200 μL of 99% methanol for 15 min. The supernatant was then discarded. Microtiter plates were air-dried for 5 min and then 200 μL of 0.4% crystal violet solution (stock solution diluted in PBS; Sigma-Aldrich) were added to each well and the plates then incubated at room

temperature for 20 min. After discarding the crystal violet solution, the wells were washed once with PBS to remove excess stain and the biomass in each well was then decolorized by adding 200 μL of 33% acetic acid for 5 min. One hundred microliters of the acetic acid solution were transferred to a new 96-well plate and the absorbance measured at 590 nm using a microplate reader (SpectraMax M3; Molecular Devices) [21].

#### *2.5. Biofilm Architecture: Confocal Laser Scanning Microscopy (CLSM) Assay*

Biofilms were formed on a polystyrene surface and treated as described above with different concentrations of micafungin (0.5–2.0 mg/L). Then, the biofilms were stained with Calcofluor white (Sigma-Aldrich) solution (5 μg/mL) for 1 h at room temperature and protected from the light [21–23]. Subsequently, the biofilms were washed twice with PBS and covered with *n*-propyl-gallate for observation using a confocal microscope (Leica TCS SP5 with OBS, Berlin, Germany). Fiji ImageJ2 software (UW-Madison LOCI, Madison, WI, USA), was used to obtain three-dimensional (3-D) reconstitutions of the biofilms [21,24]. In this way, image analysis was performed using *z*-series image stacks from five randomly chosen spots on each biofilm [21].

#### *2.6. Literature Review*

This exercise involved the compilation of available data regarding the susceptibility of the *C. haemulonii* species complex to echinocandins. The literature search was performed on 19 July 2020 using the following four databases: PubMed (https://pubmed.ncbi.nlm.nih.gov), Web of Science (https://webofknowledge.com), Google Scholar (https://scholar.google.com) and Scielo (https://scielo.org/). The term "Candida haemulonii" was added in the category "title/abstract" in the PubMed Advanced Search Builder and in the Web of Science databases, while in Google Scholar the search was conducted in the advanced search area, including the term "Candida haemulonii" and selecting the option "with the exact phrase in the title"; finally, for the Scielo database, we only used the search term "Candida haemulonii" in the general search. Papers available in English and published after the reclassification of the *C. haemulonii* complex by Cendejas-Bueno et al. [5] were selected. Subsequently, the list of results from each database was exported to the EndNote® software (version X1), using the "Output Records" tool in order to eliminate possibly duplicated references by means of the "Find Duplicates" tool. Finally, the papers were individually analyzed in order to select those that described either MIC or geometric-mean (GM)-MIC values of the *C. haemulonii* complex for echinocandins.

#### *2.7. Statistics*

All experiments were performed in triplicate, in three independent experimental sets. The results were analyzed statistically by the Analysis of Variance One-Way ANOVA (comparisons between three or more groups). All analyzes were performed using the GraphPad Prism5 program. For all analyses, *p* values of 0.05 or less were considered statistically significant.

#### **3. Results and Discussion**

#### *3.1. Susceptibility of Planktonic Cells of the C. haemulonii Species Complex to Echinocandins*

According to the breakpoints suggested in the M27S3 document published by CLSI, the planktonic cells of all clinical isolates of the *C. haemulonii* complex tested herein were considered susceptible to echinocandins, with MIC values ranging from 0.125 to 0.5 mg/L for caspofungin and 0.25–0.5 mg/L for micafungin (Table 1). For instance, a recent report described the successful use of caspofungin (MIC of ≤0.125 mg/L) in the treatment of a case of catheter-related candidemia caused by *C. haemulonii* in a pediatric patient in Mexico [25], whose fungal isolate exhibited in vitro high MICs for azoles (fluconazole MIC ≥ 256 mg/L, posaconazole ≥ 8 mg/L, itraconazole, ketoconazole and voriconazole ≥ 16 mg/L) and amphotericin B (MIC 1–2 mg/L). Some years before, a catheter-related candidemia in an adult patient hospitalized for a long period was only resolved when fluconazole treatment was replaced by caspofungin [4].


**Table 1.** MIC values of echinocandins against the *C. haemulonii* species complex studied herein.

<sup>a</sup> GM-MIC, geometric mean-minimal inhibitory concentration. <sup>b</sup> Similar results were reported in our previously published paper [6]. Journal of Antimicrobial Chemotherapy (JAC) provided the permission to reproduce this set of results.

In general, echinocandins are highly active in vitro against species comprising the *C. haemulonii* complex [7,26–29], but the existence of isolates resistant to this class of antifungals has already been reported [4,5,30]. Herein, we conducted a careful review of the literature regarding the susceptibility of the *C. haemulonii* species complex to the three clinically available echinocandins, including only papers published after the species reclassification and the creation of the *C. haemulonii* complex [5]. Using the keyword "Candida haemulonii" in the search section, 148, 63, 46 and 5 publications were located from the Web of Science, PubMed, Google Scholar and Scielo databases, respectively (Table 2). However, only a small fraction of these published papers (varying from 12.2%–28.3%) cited the in vitro susceptibility profile of the *C. haemulonii* species complex against echinocandins. In this sense, we recovered a total of 21 distinct papers that fitted our established criteria and, for these reasons, they were selected for data extraction as follows: 5 (23.8%) papers studied the three members forming the *C. haemulonii* complex, 6 (28.6%) studied only two species (*C. haemulonii* and *C. duobushaemulonii*) and 10 (47.6%) studied only one species (*C. haemulonii*, *n* = 6, *C. duobushaemulonii*, *n* = 3, *C. haemulonii* var. *vulnera*, *n* = 1). Furthermore, 13 (61.9%) papers detailed the MIC value for each isolate investigated, while the remaining studies (*n* = 8; 38.1%) only presented the geometric mean (GM)-MIC and/or the range of MIC values for the fungal isolates against the test echinocandins. Finally, 12 (57.1%) papers tested the three echinocandins, 5 (23.8%) used two and 4 (19.1%) tested only one echinocandin, with caspofungin being the most frequently evaluated.


**Table 2.** Number of publications retrieved from database searches using the term "Candida haemulonii".

The searches were conducted in PubMed (https://pubmed.ncbi.nlm.nih.gov), Web of Science (https://webofknowledge. com), Google Scholar (https://scholar.google.com/) and Scielo (https://scielo.org) on 19 July 2020. The term "Candida haemulonii" was added in the category "title/abstract" in the PubMed Advanced Search Builder and Web of Science; in Google Scholar the search was conducted in the advanced search area, including the term "Candida haemulonii" and selecting the option "with the exact phrase in the title"; in Scielo, we only searched for the term "Candida haemulonii" in the general search. Papers published after the reclassification of the *C. haemulonii* complex were included [5]. \* Papers that evaluated the susceptibility of isolates of the *C. haemulonii* species complex to echinocandins.

The results emanating from this literature review revealed that micafungin and anidulafungin appeared to be more effective than caspofungin against the three species forming the *C. haemulonii* complex (Table 3) [5,7,25,29,31–47]. In this respect, 89.8% of the isolates of *C. haemulonii* exhibited susceptibility to caspofungin, while 96.3% and 98.4% were susceptible to micafungin and anidulafungin, respectively. Regarding *C. duboushaemulonii*, 95.5% of the isolates were susceptible to caspofungin, 99.1% to anidulafungin and 100.0% to micafungin. Finally, considering the clinical isolates of *C. haemulonii* var. *vulnera*, 85.0% were susceptible to caspofungin, 91.7% to micafungin and 97.1% to anidulafungin. Indeed, the MIC frequency distribution demonstrated that the modal MIC of echinocandins against the *C. haemulonii* complex was ≤0.12 mg/L in almost all cases (Table 4).

**Table 3.** Literature compilation regarding the distribution (%) of the susceptible (S) and non-susceptible (NS) isolates belonging to the *C. haemulonii* complex against echinocandins described in published papers available until 19 July 2020.


\* Antifungal susceptibility testing was interpreted according to the document M27-S3 published by CLSI; *n*, number of fungal isolates; the references used to construct this table were [5,7,25,29,31–47].

Comparing the GM-MIC values of our clinical isolates (Table 1) with those compiled from the literature reports (for these comparisons, we used the arithmetic mean of the GM-MIC values of the selected works, as summarized in Table 5), we observed that the GM-MIC values of caspofungin for our isolates of *C. haemulonii*, *C. duobushaemulonii* and *C. haemulonii* var. *vulnera* were higher than those reported in the literature (0.33 mg/L versus 0.18 mg/L for *C. haemulonii*, 0.18 mg/L versus 0.11 mg/L, for *C. duobushaemulonii* and 0.32 mg/L versus 0.21 mg/L for *C. haemulonii* var. *vulnera*). Similarly, GM-MIC values for micafungin calculated from the literature reports were lower than ours (0.18 mg/L versus 0.33 mg/L for *C. haemulonii*, 0.17 mg/L versus 0.30 mg/L for *C. duobushaemulonii*, and 0.13 mg/L versus 0.25 mg/L for *C. haemulonii* var. *vulnera*). Finally, based on the analysis of the literature data, anidulafungin also produced low GM-MIC values for the three fungal species of the

*C. haemulonii* complex (GM-MICs of 0.16, 0.32 and 0.06 mg/L for *C. haemulonii*, *C. duobuhaemulonii* and *C. haemulonii* var. *vulnera*, respectively).

**Drug <sup>a</sup>** *Species* **MIC (mg**/**L) MIC50 <sup>b</sup> MIC90 <sup>c</sup>** ≤**0.015 0.03 0.06 0.12 0.25 0.5 1 2 4 8 16** >**16 Range CAS** *Ch* 19 17 14 12 6 1 1 1 14 0.03–>16 0.12 >16 *Cd* 3 14 18 20 9 4 1 11 3 ≤0.015–>16 0.12 0.5 *Chv* 254 3 0.12–>16 0.25 >16 **MCF** *Ch* 8 12 28 841 4 ≤0.015–>16 0.06 0.25 *Cd* 2 12 36 12 3 1 0.06–0.5 0.06 0.12 *Chv* 1 4 1 0.06–>16 0.12 0.12 **ANF** *Ch* 27 14 19 10 3 1 1 2 ≤0.015–>16 0.03 0.12 *Cd* 11 8 17 16 15 4 3 1 1 ≤0.015–4 0.12 0.5 *Chv* 8 14 1 ≤0.015–>16 ≤0.015 0.06

**Table 4.** MIC distribution of *C. haemulonii* complex isolates obtained from the literature review against the three echinocandins.

<sup>a</sup> CAS, caspofungin; MCF, micafungin; ANF, anidulafungin; <sup>b</sup> MIC50, MIC at which 50% of isolates were inhibited; <sup>c</sup> MIC90, MIC at which 90% of isolates were inhibited; Modal MICs are indicated with underlined numbers; MIC values of <0.03 were allocated as ≤0.015; Clinical and Laboratory Standards Institute (CLSI) document M27S3 suggests the following breakpoints for echinocandins against *Candida* spp.: susceptible ≤ 2 mg/L and non-susceptible > 2 mg/L; the references used to construct this table were [5,29,31,34,35,38–42,44,46,47].


**Table 5.** Literature review on the antifungal susceptibility of different isolates of the *C. haemulonii* complex to echinocandins.


**Table 5.** *Cont.*

\* GM-MIC, geometric mean of the minimal inhibitory concentrations expressed in mg/L; • Values of GM-MIC obtained directly from the papers; ◦ Values of GM-MIC calculated by us from the MIC values for each isolate mentioned in the articles; *Ch*, *C. haemulonii*; *Cd*, *C. duobushaemulonii*; *Chv*, *C. haemulonii* var. *vulnera*; *n*, number of isolates studied; arithmetic mean of overall GM-MIC calculated from the GM-MIC of the different papers; ND, not determined; -, no isolates were tested.

In summary, the majority of literature reported GM-MIC concentration values of <0.5 mg/L for the three echinocandins against the *C. haemulonii* species complex. Nevertheless, two works warranted specific attention: Cendejas-Bueno et al. [5], in which the GM-MIC values for caspofungin for the three members of the *C. haemulonii* complex were disproportionately high in comparison to our present results and those given in the other literature publications; and Isla et al. [36], in which the GM-MIC value obtained for caspofungin against the *C. duobushaemulonii* isolates was considerably higher (Table 5). A possible explanation for the high MIC values found in the aforementioned papers is the possible occurrence of paradoxical growth effect (also known as the Eagle effect), that is characterized by reduced activity of the antifungal agents at high concentrations. In fact, Cendejas-Bueno et al. [5] stressed this discussion in their study, but in a superficial way. A recent study conducted with 106 clinical isolates of *C. auris* demonstrated that the vast majority of isolates were susceptible to the echinocandins; however, they exhibited different intensities of paradoxical growth effect in the presence of caspofungin, whilst four isolates were resistant to echinocandins and had a mutation in hot spot region 1 of the *FKS* gene [48]. Interestingly, those isolates presenting paradoxical growth effect were susceptible to caspofungin at doses used in human treatment, while those with *FKS1* mutation were still resistant in a murine model of invasive candidiasis, demonstrating that only the isolates with the mutations display in vivo echinocandin resistance [48].

#### *3.2. E*ff*ects of Echinocandins on the Biofilm Formed by C. haemulonii Species Complex*

In order to evaluate the effects of echinocandins (caspofungin and micafungin) on the viability and biomass of the biofilms formed by the clinical isolates of the *C. haemulonii* complex, the mature biofilms were firstly incubated with different concentrations of the antifungals and then analyzed. The metabolic activity of viable fungal cells was assessed by their ability to reduce XTT to formazan, whilst the decrease in biofilm biomass was measured spectroscopically by looking at the incorporation of crystal violet into methanol-fixed, non-viable cells (Figures 1 and 2). In general, the test echinocandins were found to be more efficient at reducing cell viability than decreasing the biomass of the *C. haemulonii* complex biofilms.

The decrease of both viability and biomass parameters by caspofungin was isolate-dependent. At the lowest concentration used (0.25 mg/L) this echinocandin caused a statistically significant reduction in the viability of all of the fungal cells tested (*p* < 0.05; One-way ANOVA analysis of variance, Dunnett's multiple comparison test), varying from 30–80% among the different isolates (Figure 1). However, caspofungin was unable to reduce the biomass of some of the *C. haemulonii* isolates (LIP*Ch*2, LIP*Ch*3 and LIP*Ch*4) even at the highest concentration used. Nevertheless, for the remaining fungal isolates the drug caused a biomass reduction of up to 60% (mainly against the *C. duobushaemulonii*

isolates) (Figure 2). The isolates LIP*Ch*2 (*C. haemulonii*), LIP*Ch*1 (*C. duobushaemulonii*) and LIP*Ch*5 (*C. haemulonii* var. *vulnera*) were less susceptible to caspofungin at the higher concentrations (Figure 1).

**Figure 1.** Cell viability of biofilms formed by clinical isolates comprising the *C. haemulonii* complex exposed to different concentrations of echinocandins (caspofungin and micafungin). The results were assessed spectroscopically (492 nm) by XTT reduction and expressed as the mean of metabolic activity percentages compared to untreated biofilms (control), which correspond to 100%. The graphs exhibit the mean ± standard deviation of three independent experiments. The dashed boxes represent the concentrations of echinocandins that caused statistically significant reduction of cell viability in relation to the respective control (*p* < 0.05; One-way ANOVA analysis of variance, Dunnett's multiple comparison test).

**Figure 2.** Biomass of biofilms formed by clinical isolates comprising the *C. haemulonii* species complex exposed to different concentrations of echinocandins (caspofungin and micafungin). The amount of crystal violet incorporated by the cells was assessed spectroscopically (absorbance at 590 nm) and the results expressed as the mean of biomass percentages compared to untreated biofilms (control), which correspond to 100%. The graphs show the mean ± standard deviation of three independent experiments. The dashed boxes represent the concentrations of echinocandins that caused a statistically significant reduction in biomass in relation to the respective control (*p* < 0.05; One-way ANOVA analysis of variance, Dunnett's multiple comparison test).

Micafungin proved to be more effective than caspofungin at disturbing both biofilm viability and biomass. A decrease in biofilm viability of up to 60% was seen among most of the clinical isolates, especially against *C. duobushaemulonii* and *C. haemulonii* var. *vulnera* (Figure 1). Unlike caspofungin, micafungin showed a decrease of up to 60% on the biofilm biomass of *C. haemulonii* isolates, with the exception of isolate LIP*Ch*4, which forms a very dense and robust biofilm (Figure 2). For the *C. duobushaemulonii* and *C. haemulonii* var. *vulnera* isolates, micafungin reduced biomass in the range 20–60% (Figure 2). In summary, the lowest concentration of micafungin used was able to significantly reduce the cell viability and the biomasses of biofilms formed by all of the test isolates, expect for the biomass of one isolate.

The determination of MBEC, which was defined as the lowest antifungal concentration able to reduce the biofilm viability in 50% [19], revealed that the biofilms of all isolates remained susceptible to echinocandins, with the exception of the isolate LIP*Ch*4 of *C. haemulonii* (Table 6). This fact could be explained by the ability of the isolate LIP*Ch*4 to form very robust biofilm on polystyrene in comparison with the other isolates [18,21], hampering the action of echinocandins due to the high amount of fungal cells-forming the biofilm architecture as well as due to the high production of extracellular matrix that can block the antifungal penetration into the biofilm structure.


**Table 6.** Minimal biofilm eradication concentration (MBEC) to echinocandins against*C. haemulonii* complex.

As micafungin was more active than caspofungin against the mature biofilms formed by the *C. haemulonii* species complex it was chosen for further studies. In order to verify the 3-D organization of the biofilms following exposure to micafungin two isolates of *C. haemulonii* were selected: LIP*Ch*3, to represent the isolates having susceptible biofilms, and LIP*Ch*4, to represent isolates forming resistant biofilms. CLSM analysis was conducted using Calcofluor white, which binds to the chitin in the fungal cell wall, to evidence the biofilm biomass. The CLSM analysis corroborated the results observed by crystal violet approach, with the lowest antifungal concentration used causing a drastic reduction in the biofilm biomass of LIP*Ch*3, whilst even the highest concentration of micafungin failed to affect the biofilm formed by LIP*Ch*4 (Figure 3).

Until now, no information has been available in the literature regarding the activity of conventional antifungal agents against the biofilm formed by the *C. haemulonii* species complex. A recent study conducted with *C. auris*, which belongs to the *C. haemulonii* clade, showed that, despite the susceptibility of planktonic cells to echinocandins and amphotericin B, the biofilms were not vulnerable, exhibiting MBECs which were 512-fold higher than their planktonic MIC counterparts [19]. Actually, the biofilm formed by *C. auris* is not as robust as those arising from *C. albicans* and *C. glabrata*, but its tolerance to the major classes of antifungal agents is notable, especially for amphotericin B and micafungin, which are the recommended antifungal therapeutics for infections caused by *C. albicans* biofilms [49]. The antifungal tolerance of the *C. auris* biofilm has been shown to be phase-dependent, with the mature biofilms resistant to the three available antifungal drug classes [50]. On the other hand, micafungin has been shown to be effective against both planktonic and biofilm-forming *C. albicans* cells, while its effectiveness against *C. parapsilosis* was considered to be moderate [51]. Additionally, micafungin concentrations >2 mg/L prevented the regrowth of *Candida* biofilm cells [51]. Regarding the *C. parapsilosis* complex, caspofungin was more active against biofilms of *C. orthopsilosis* than *C. parapsilosis* sensu strictu, with 20% and 86% of isolates resistant to this antifungal, respectively, suggesting that a treatment of catheter-related candidemia caused by *C. orthopsilosis* with caspofungin would be more effective than

against *C. parapsilosis* sensu strictu [52]. A study, conducted with five different *Candida* species recovered from cases of bloodstream infections demonstrated both species-specific and drug-specific differences in *Candida* biofilms regarding their susceptibility to echinocandins [53]. In this sense, while *C. albicans* and *C. krusei* biofilms were susceptible to the three clinically available echinocandins, *C. lusitaniae*, *C. guilliermondii* and *C. parapsilosis* were quite resistant to them [53]. In addition, micafungin seemed to be the most effective echinocandin against *C. parapsilosis* biofilms, presenting lower MBECs against this *Candida* species in comparison to caspofungin and anidulafungin [53]. These observations reinforce the need to determine the correct identification of the actual fungal species causing the candidiasis infection and, further, to assess its antifungal susceptibility profile against both planktonic and biofilm-forming cells in order to choose the best therapeutic option for each case.

**Figure 3.** Representative confocal laser scanning microscopy (CLSM) images of the biofilms formed by *C. haemulonii* on a polystyrene surface. Yeasts (200 μL containing 106 cells) were placed to interact with the

polystyrene for 48 h at 37 ◦C. Subsequently, the supernatant fluids were removed and washed with PBS, and 200 μL of RPMI 1640 medium containing different concentrations of micafungin were added. The biofilms were incubated at 37 ◦C for an additional 48 h. Afterwards, the supernatant fluids were carefully removed again, and the wells were washed twice with PBS to remove non-adherent cells. Finally, the biofilms were stained with Calcofluor white in order to evidence the fungal biomass. The panels on the left represent the top view images of the fungal biofilms visualized by Confocal Laser Scanning Microscopy (CLSM) (bars represent 5 μm). The graphs on the right represent the three-dimensional reconstruction of the biofilms formed. The isolate LIP*Ch*3 of *C. haemulonii* (**A**) was chosen to represent susceptible biofilms, while the isolate LIP*Ch*4 of *C. haemulonii* (**B**) represents resistant biofilms.

Furthermore, we observed that one isolate of each species forming the *C. haemulonii* complex showed a smaller reduction in cell viability when incubated in the presence of higher concentrations of the echinocandins. This phenomenon is called paradoxical growth, and it corresponds to the decreased sensitivity to echinocandins in the presence of concentrations higher than the MIC values. To date, the evidence strongly suggests that this paradoxical effect is more commonly associated with caspofungin than either micafungin or anidulafungin [54]. This effect has already been documented for biofilms formed by other *Candida* species, such as *C. albicans* [53,55], *C. parapsilosis* [53], *C. tropicalis* [55] and *C. dubliniensis* [56].

To finalize, we recognize some of the limitations associated with the present study, such as the limited number of isolates used and the exclusion of anidulafungin. The experiments were conducted with only 12 clinical isolates of the *C. haemulonii* complex due to the difficulties in obtaining more isolates, since it is quite a rare fungal complex. Additionally, we tested only two of the three echinocandins currently in clinical use, and this was because at the time the experiments were conducted anidulafungin was not available for scientific research purposes.

#### **4. Conclusions**

In addition to their own clinical conditions, hospitalized patients are at constant risk of acquiring contagions associated with the hospital environment. Biofilm-related *Candida* infections represent an important and worrisome threat to these patients, and there is a limited number of available antifungal agents of sufficient potency to break down these highly resistant structures. In this sense, echinocandins are considered highly active against various *Candida* species and the results presented herein reinforce the potential of echinocandins to treat biofilm-related infections caused by the emergent and multidrug-resistant species comprising the *C. haemulonii* complex.

**Author Contributions:** All authors conceived and designed the experiments. L.S.R. performed the experiments. All authors analyzed the data. M.H.B. and A.L.S.S. contributed reagents/materials/analysis tools. All authors wrote and revised the paper. All authors contributed to the research and approved the final version of the manuscript. All authors agree to be accountable for all aspects of the work. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES—Financial code 001).

**Acknowledgments:** The authors would like to thank Denise Rocha de Souza (UFRJ) for technical assistance in the experiments and Grasiella Matioszek (UFRJ) for confocal analyses. The authors would like to thank Malachy McCann (Department of Chemistry at the National University of Ireland Maynooth—NUIM, Ireland) for his valuable contribution to the critical review and editing of English.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Fungal Quorum-Sensing Molecules: A Review of Their Antifungal E**ff**ect against** *Candida* **Biofilms**

**Renátó Kovács 1,2,\* and László Majoros <sup>1</sup>**


Received: 4 June 2020; Accepted: 30 June 2020; Published: 2 July 2020

**Abstract:** The number of effective therapeutic strategies against biofilms is limited; development of novel therapies is urgently needed to treat a variety of biofilm-associated infections. Quorum sensing is a special form of microbial cell-to-cell communication that is responsible for the release of numerous extracellular molecules, whose concentration is proportional with cell density. *Candida*-secreted quorum-sensing molecules (i.e., farnesol and tyrosol) have a pivotal role in morphogenesis, biofilm formation, and virulence. Farnesol can mediate the hyphae-to-yeast transition, while tyrosol has the opposite effect of inducing transition from the yeast to hyphal form. A number of questions regarding *Candida* quorum sensing remain to be addressed; nevertheless, the literature shows that farnesol and tyrosol possess remarkable antifungal and anti-biofilm effect at supraphysiological concentration. Furthermore, previous in vitro and in vivo data suggest that they may have a potent adjuvant effect in combination with certain traditional antifungal agents. This review discusses the most promising farnesol- and tyrosol-based in vitro and in vivo results, which may be a foundation for future development of novel therapeutic strategies to combat *Candida* biofilms.

**Keywords:** *Candida*; farnesol; tyrosol; biofilm; therapy; combination

#### **1. Introduction**

It has been estimated that there are 2.2 to 3.8 million fungal species worldwide; however, approximately 300 species have been described to cause human disease [1]. *Candida* species are among the most common human fungal pathogens. The annual incidence rate of *Candida*-associated bloodstream infections ranged from 9.5 to 14.4 per 100,000 in the United States of America [2]. This value ranged from 1.4 to 5.7 per 100,000 in Europe, depending on the country [3]. In the last two decades, the prevalence of resistant fungal infections has been steadily increasing due to the widespread use of antifungals in agriculture and veterinary and human medicine [4,5]. Global warming and anthropogenic effects have resulted in the emergence of previously little-known, potentially multi-resistant fungal pathogens in clinical practice, such as *Candida auris*, azole-resistant *Aspergillus* spp., or *Lomentospora prolificans*. These emerging pathogens have caused further challenges for therapy [6,7].

Several fungal species can switch their morphology from yeast to hyphal or pseudohyphal forms, which is coupled with biofilm formation and plays a pivotal role both in fungal virulence and in resistance to antifungals [8–10]. The increased number of biofilm-associated infections is exacerbated by a paucity of antifungal agents or therapeutic strategies in development that have unique mechanisms of action or possess alternative approaches, respectively [11]. Currently, the most promising antifungal agents are already in Phase 3 including ibrexafungerp [12], rezafungin [13], super bioavailable itraconazole [14], and VT-1161 [15]. Recently investigated alternative therapeutic approaches involve high-dose therapy with available antifungal agents [16–18], antifungal lock therapy [19], and combination-based therapies [20,21]. Based on in vitro and in vivo data, echinocandins and amphotericin B solutions are the most promising combination-based and/or antifungal lock strategies [19]. Further innovative therapeutic approaches may be the natural products-based treatments [22,23]. One of the more well-studied compounds is carbohydrate-derived fulvic acid as a heat stable colloidal material, which has an inhibitory effect on *Candida* and bacterial biofilm formation [24]. Moreover, a further alternative approach is the treatments disrupting quorum sensing. The usage of quorum sensing molecules at supraphysiological concentration may adversely influence the cell-to-cell communication in biofilms [25–27]. In addition, the quorum-sensing system can be inactivated, which is generally known as quorum quenching. Quorum quenching can be triggered by inhibiting the production of quorum sensing molecules, their detection by receptors or their degradation [28].

In this review, a detailed overview is provided of the recent status of quorum-sensing molecule-based therapeutic approaches and their potential future perspectives against *Candida* biofilms.

#### **2. The Medical Importance of** *Candida* **Biofilms**

Despite their importance, *Candida* biofilms remain a relatively underappreciated and understudied area. Therefore, effective therapeutic strategies against these sessile communities remain scarce. Biofilms are usually found in medical devices such as joint prostheses, pacemakers, urinary and central venous catheters, dentures, and mechanical heart valves, hindering the eradication of *Candida* infections [10]. In addition, several chronic *Candida*-related diseases are also associated with biofilm development [29]. Biofilm formation on the vaginal mucosa has been observed in in vivo models of vulvovaginal candidiasis [30]. Oral- and oesophageal mucosae-associated biofilms are a very important contributor to oral diseases caused by *Candida* species; gastrointestinal and urogenital tracts are also common sites of *Candida*-associated opportunistic infections [31]. *Candida* is one of the most commonly identified fungal genera in wounds whose environment can also promote the formation of biofilms [32]. A series of recent studies has indicated that strains defective in hyphal formation display significantly milder symptoms, highlighting the role of biofilm formation in pathogenesis of these chronic or recurrent infections [30,33].

These sessile communities exhibit five- to eightfold higher resistance to all licenced antifungal drugs when compared to their planktonic counterparts [10]. This high rate of resistance can be explained by the increased metabolic activity of cells in the early development phase of biofilm formation [10]. On the other hand, dormant, non-proliferating persister cells have also been observed, especially in mature biofilms, that have demonstrated high tolerance to antifungals [34]. Furthermore, the various *Candida* species can produce dense extracellular polymeric substances which serve as a solid barrier to prevent the diffusion of antifungal drugs and account for almost 90% of the biofilm dry mass [10]. As has been previously reported in the literature, sessile *Candida* communities exhibit an altered gene expression profile, including the upregulation of *CDR* and *MDR* genes which encode azole resistance transporter proteins, and pose further challenges for treatment [35].

To date, there is no definitive therapy against *Candida* biofilms; nevertheless, there are several promising in vitro, in vivo and clinical results. The increasing number of resistant *Candida* species and isolates highlight the need for new molecules with new targets. Alternative therapeutic approaches against multidrug-resistant fungal biofilms may be the result of a combination of traditional antifungal agents with quorum-sensing molecules [36].

#### **3. Fungal Quorum Sensing**

A major mechanism of microbial communication is a population density-dependent stimulus-response system called quorum sensing. This process occurs by the continuous release and monitoring of low molecular weight hormone-like secreted molecules (quorum-sensing molecules), which are not elementary in the central metabolism but have a variety of biological activities. The concentration of these quorum-sensing molecules is proportional with the size of population;

after reaching a critical threshold, a response is triggered leading to the coordinated expression or repression of quorum sensing-related target genes [37].

In the fungal kingdom, quorum sensing was a relatively unknown phenomenon until Hornby et al. (2001) described the effect of the isoprenoid farnesol on *Candida albicans* morphogenesis; this opened a new branch of science focusing on fungal quorum sensing [38]. At the same time, quorum sensing has been already reported in *Aspergillus* spp. [39] and *Penicillium* spp. [40]. To date, four main quorum-sensing molecules were described including farnesol, tyrosol, phenylethanol, and tryptophol, which have a remarkable effect on the regulation of morphogenesis (yeast to hyphae transition and vice versa), initiation of fungal apoptosis, and virulence [41].

Recently, several authors reported that certain quorum-sensing molecules may generate oxidative stress, especially at supraphysiological concentrations, which may have an antifungal effect [42–45]. The majority of data concerning fungal quorum sensing molecule-related therapeutic potential derived from *C. albicans* experiments, and these results cannot be always directly extrapolated to non-*albicans* species. Recently, the number of studies dealing with the effect of quorum-sensing molecules on non-*albicans* species has steadily increased, supporting the comprehensive understanding of the in vitro and in vivo antifungal effects exerted by these molecules.

#### **4. Farnesol**

#### *4.1. Physiological E*ff*ect of Farnesol in Candida Species*

Farnesol (3,7,11-trimethyl-2,6,10-dodecatriene-1-ol) was the first described *Candida*-derived quorum sensing molecule; it is released in *C. albicans* as a side product of the sterol synthetic pathway by dephosphorylation of farnesol pyrophosphate [38,46]. It is an acyclic sesquiterpene heat-stable molecule, which is produced primarily under aerobic conditions and it is unaffected by extreme pH and the type of carbon or nitrogen source [38,47]. Generally, the farnesol concentration is proportional to the colony-forming unit number [38]. Under physiological conditions, *C. albicans* isolates secrete a farnesol concentration with a mean of 35.6 μM (range: 13.7 to 58.5 μM) [48]. This concentration was 35 times higher than that secreted by non-*albicans* species, with the exception of *Candida dubliniensis*, which has demonstrated a concentration of 8.3 μM (range: 6.0 to 17.5 μM). All other non-albicans species excreted significantly lower farnesol concentrations, ranging from 0.4 to 1 μM [48]. These differences in excretion may be explained by the species-specific characteristics in sterol synthesis [49].

Based on a cDNA microarray analysis, a total of 274 genes were identified as responsive in *C. albicans*, with 104 genes up-regulated and 170 genes down-regulated [50]. Farnesol has an ability to influence *Candida* morphology, biofilm formation, drug efflux pump expression, apoptosis regulation, phagocytic response, surface hydrophobicity, iron metabolism, and heat-shock-related pathways [50–54]. One of the most prominent farnesol-associated effects is the induction of hypha-to-yeast transition and the inhibition of biofilm formation in various *Candida* species. It should be emphasized that 150-fold more farnesol is needed to block germ-tube formation in the presence of 10% serum, showing that it can bind to serum proteins at a high rate [55,56].

In view of this diverse role, it is not surprising that this compound influences several central signalling pathways in different *Candida* species. One of the best-studied farnesol-related pathways is the Ras1-cAMP-PKA cascade, where farnesol binds to the cyclase domain of the adenylyl cyclase *Cyr1*, influencing the level of intracellular cAMP [57]. Moreover, farnesol induces the cleavage of the small GTPase Ras1, resulting in a soluble Ras1; soluble Ras1 is a weak activator of *Cyr1* and supports the formation of yeast cells [58]. Furthermore, farnesol can directly inhibit the cAMP signalling pathway, supporting the hypha-to-yeast transition [59]. It is noteworthy that farnesol exposure stabilizes the *Nrg1* protein, which is the negative regulator of filamentation [60]. While farnesol was described first in *C. albicans*, it can inhibit filamentation and growth in other fungal species [27,61], including *Saccharomyces cerevisiae* [62], *Aspergillus niger* [63],

#### *Aspergillus flavus* [64], *Aspergillus nidulans* [65], *Penicillium expansum* [66], *Fusarium graminearum* [67], and *Paracoccidioides brasiliensis* [68].

Regarding reactive oxygen species production, the supraphysiological farnesol concentrations (200–300 μM) are stressful for most fungi, while the physiological concentrations (30–40 μM) protect them from stress [57]. In addition to the farnesol-related effect on growth in the case of different microbes, the molecule also has a relevant immunomodulator effect [57,69]. Farnesol can stimulate both macrophage chemokine synthesis or macrophage recruitment, and trigger activation of neutrophil granulocytes and monocytes. Farnesol exposure also influences the differentiation of monocytes into dendritic cells [57,69].

Farnesol has been reported to induce cell growth inhibition and/or apoptosis in tumor cells where the observed IC50 values varied widely for different tumor types and different cell lines [70]. Farnesol caused 100% cell death at>120 μM in A549 and H460 lung cancer cells [71]. Scheper et al. (2008) observed an IC50 value of 30 to 60 μM for farnesol on the primary human tongue squamous cell carcinoma cell lines (OSCC9, OSCC 25) [70]. Nagy et al. (2020) evaluated 10 μM, 50 μM, 150 μM, and 300 μM farnesol concentrations in terms of toxicity to the Caco-2 cell line, where no toxicity was observed with any concentration tested [45].

#### *4.2. Antimicrobial Activity of Farnesol*

At physiological concentrations, farnesol has no significant effect on *Candida* cells that have already begun hyphae development or biofilm formation [25,38]. However, prior results suggest that farnesol can cause biofilm degradation at supraphysiological concentrations, suggesting the potential use of this compound in biofilm-associated infections [36]. In addition, several authors have published studies demonstrating contribution of farnesol to reduced azole resistance of *Candida* cells, including in biofilms [72]. This phenomenon can be explained by the modulation of *Cdr1* efflux pumps, reactive oxygen species production, or changes in glutathione homeostasis [38,61,72]. Furthermore, farnesol has an effect on genes connected to ergosterol synthesis [46]. Dižová et al. (2018) observed that the presence of 200 μM farnesol down-regulated the *ERG20*, *ERG11* and *ERG9* genes. However, this farnesol concentration supplemented with 0.5 mg/L fluconazole restored the original expression level of *ERG20* and *ERG11*. Interestingly, the physiological farnesol concentration (~30 μM) only slightly influences the expression of these genes in 48 h-old biofilms [73]. Chen et al. (2018) reported that *CYR1* and *PDE2* regulate resistance mechanisms against various antifungals in *C. albicans* biofilms. However, farnesol can diminish the resistance of *C. albicans* biofilms by regulating the expression of the gene *CYR1* and *PDE2* [74]. Yu et al. (2012) observed that the sterol biosynthetic pathway may contribute to the inhibitory effects of farnesol, as the transcription levels of the *ERG11*, *ERG25*, *ERG6*, *ERG3*, and *ERG1* genes decreased following farnesol exposure [75]. Jabra-Rizk et al. (2006) showed that farnesol concentrations of 30–50 mM decrease the fluconazole MICs for *C. albicans* and *C. dubliniensis* from resistant values to a susceptible dose-dependent range, while concentrations of 100–300 mM resulted in fluconazole susceptibility [76].

One of the first major breakthroughs in combination-based experiments with farnesol and antifungals was published by Katragkou et al. (2015), who found a significant synergy against *C. albicans* 48 h-old biofilms between fluconazole, amphotericin B, and micafungin in the presence of farnesol [26]. The highest synergistic effect was observed in the case of micafungin combined with farnesol using fractional inhibitory concentration index determination and Bliss independence analysis. Based on the Bliss model, the observed effects were 39–52% higher compared to the expected efficacy if the drugs had been acting independently [26]. It should be noted that synergism was observed only in the case of farnesol/micafungin and farnesol/fluconazole based on calculated fractional inhibitory concentration indices, suggesting the usage of multiple analytical approaches for investigation of drug-drug interaction [26].

Regarding non-albicans species, Kovács et al. (2016) showed that farnesol consistently enhanced the activity of caspofungin and micafungin, as concordantly shown in two independent experimental settings (chequerboard dilution and time–kill experiments) [27]. Fernández-Rivero et al. (2017) reported that a supraphysiological farnesol concentration (300 μM) improved the activity of amphotericin B against *Candida tropicalis* biofilms but did not affect anidulafungin [77]. Two recent studies by Nagy et al. concluded that farnesol significantly enhanced the activity of echinocandins and triazoles against one-day-old *C. auris* biofilms in vitro, suggesting an alternative approach to overcome the previously well-documented azole and echinocandin resistance of *C. auris* biofilms [45,78].

Animal experiments with farnesol raised several questions in terms of in vivo applicability of this compound. In one of the first in vivo studies, Navarathna et al. (2007) concluded that the physiological farnesol production may play a pivotal role as a virulence factor in fungal pathogenesis; furthermore, exogenous oral and intraperitoneal farnesol administration (20 mM) enhances the mortality of mice in their systemic mouse model [79]. Contrary to these results, Hisajima et al. (2008) observed a protective effect against *C. albicans* in their oral candidiasis mouse model [80]. It should be noted that there was a 1000-fold difference between the administered farnesol dosages (9 μM/mouse) in the experiments of Hisajima et al. (2008) [80] compared to experiments performed by Navarathna et al. (2007) (20 mM/mouse) [79]. In addition, they reported a potential gastrointestinal tract-related farnesol effect including moderate bodyweight reduction and reduced *Candida* faeces burden [80]. A cocktail of *Candida*-derived regulatory alcohols combined with nanomolar amounts of farnesol was reported to have a similar protective effect by Martins et al. (2012) in their murine model of disseminated candidiasis [81]. Bozó et al. (2016) did not find a farnesol-related protective effect against vaginal *C. albicans* infection [82], in contrast to the findings of Hisajima et al. (2008) [80]. However, both administered farnesol regimens enhanced the activity of 5 mg/kg daily fluconazole treatment against fluconazole-resistant *C. albicans* strain [82]. Similar fluconazole resistance reversion was observed in the case of planktonic cells by Jabra-Rizk et al. (2006) [76] and Cordeiro et al. (2013) [83]. Fernandes Costa et al. (2019) used nanoparticles containing farnesol alone or in combination with miconazole; nanoparticles containing farnesol inhibited yeast-to-hyphae transition at concentrations greater than or equal to 240 μM [84]. In addition, chitosan nanoparticles containing miconazole (33 mg/L) and farnesol (2.1 mM) inhibited fungal proliferation and decreased the pathogenicity of mouse vulvovaginitis infection [84]. Nagy et al. (2020) demonstrated that a daily treatment with 75 μM farnesol decreased the *C. auris* kidney burden in their immunocompromised systemic mouse model, especially when inocula was pre-exposed to farnesol [45].

The farnesol-exerted antifungal activity can be explained by the higher level of reactive oxygen species, especially in the case of non-*albicans* species [43,45]. Furthermore, farnesol has an amphiphilic property which allows for its integration into cell membranes, influencing membrane fluidity and integrity. In the case of *Candida parapsilosis* and *C. dubliniensis*, farnesol affected the cellular polarization and membrane permeability [61,76,85]. These observations can help further elucidate the antifungal effect.

Farnesol has a remarkable antibacterial effect alone or in combination with traditional antibacterial agents as demonstrated by in vitro investigations. Jabra-Rizk et al. (2006) observed that farnesol treatment (100 μM) increases the activity of gentamicin against *Staphylococcus aureus* biofilms [86]. Gomes et al. (2009) showed that farnesol exposure (300 μM) produced a relatively long post-antimicrobial effect (>8 h) against *Staphylococcus epidermidis* [87], while Pammi et al. (2011) observed that farnesol exposure at a concentration of 500 μM significantly inhibited the *S. epidermidis* biofilm formation in vitro [88]. A clear synergistic interaction was observed between farnesol and nafcillin or vancomycin against *S. epidermidis* sessile cells [88]. Additionally, it potentiates the activity of beta-lactam antibiotics against antibiotic-resistant bacterium species [89]. Castelo-Branco et al. (2012) showed a potent antimicrobial effect exerted by exogenous farnesol exposure against mature *Burkholderia pseudomallei* biofilms [90]. Additionally, it increased the activity of amoxicillin, ceftazidime, doxycycline, and sulfamethoxazole-trimethoprim, which are routinely administered for the treatment of melioidoses [91]. Farnesol also had a synergizing effect against ciprofloxacin-resistant *Pseudomonas aeruginosa* biofilms when used in combination with ciprofloxacin [92]. In vivo data also supports the

antibacterial efficacy of farnesol. It has been observed that 6.7 mM farnesol treatment significantly decreased the *S. epidermidis* associated catheter infection and systemic dissemination [88].

Based on several studies, farnesol has a remarkable effect in *Candida*-bacterium mixed biofilms. *C. albicans*-derived farnesol has also been shown to have an effect on the response of *S. aureus* to antibiotics in mixed species biofilms. Farnesol exposure results in a significant decrease in staphyloxantin, which is an important virulence factor of this bacterium [42]. Cern ˇ áková et al. (2018) showed that 200 μM farnesol has an inhibitory effect on *C. albicans* growth in mixed-species biofilms with *Streptococcus mutans* [93]. Cugini et al. (2010) examined the *C. albicans*-*P. aeruginosa* mixed species biofilms, where the *C. albicans*-derived farnesol enhanced *P. aeruginosa* quinolone signal production in a LasR-defective strain [94].

#### **5. Tyrosol**

#### *5.1. Physiological E*ff*ect of Tyrosol in Candida Species*

Tyrosol (2-(4-hydoxyphenyl)-ethanol) is a tyrosine-derived molecule which is synthetized via either tyramine or 4-hydroxyphenylacetaldehyde [95,96]. In the case of *C. albicans*, it is released into the growth medium continuously during the exponential growth phase and is capable of decreasing the duration of the lag phase before cells begin germination. The accumulation of tyrosol in the culture medium is proportional to the rise of fungal cell number. While the molecule stimulates filamentation, it exclusively promotes germ tube formation in conditions that normally induce these physiological processes [95,96]. Tyrosol exposure influences cell cycle regulation, DNA replication, and chromosome segregation in *C. albicans* [95]. Additionally, it was shown to have an inhibitory effect on neutrophil granulocytes by interfering with the oxidative stress response of these phagocytes [97,98]. Significantly more tyrosol was secreted by *C. albicans* (range: 21.01 <sup>±</sup> 0.76 to 53.40 <sup>±</sup> 1.73 <sup>μ</sup>M/1.6 <sup>×</sup> <sup>10</sup>7–5.3 <sup>×</sup> 107 cells/mL) and *C. tropicalis* (range: 41.21 <sup>±</sup> 1.21 to 48.63 <sup>±</sup> 3.83 <sup>μ</sup>M/2.6 <sup>×</sup> 107–2.7 <sup>×</sup> 107 cells/mL) than by *Candida glabrata* (range: 1.3 <sup>±</sup> 0.17 to 3.26 <sup>±</sup> 0.33 <sup>μ</sup>M/2.7 <sup>×</sup> 107–5.5 <sup>×</sup> 10<sup>7</sup> cells/mL) or *C. parapsilosis* (range: 1.59 <sup>±</sup> 0.29 to 3.04 <sup>±</sup> 0.43 <sup>μ</sup>M/1.7 <sup>×</sup> <sup>10</sup>7–2.3 <sup>×</sup> 107 cells/mL), suggesting a possible link with virulence [99]. Tyrosol plays a pivotal role in biofilm production, where it can stimulate hypha production of *C. albicans*, especially between two and six hours of biofilm development. *C. albicans* biofilms released at least 50% more tyrosol when compared to planktonic cells [96].

Regarding non-*albicans* species, tyrosol has been recognized as inducing the biofilm-forming ability of *C. auris*to grow as yeast or pseudohyphae [96]. Based on RNA-Seq analysis, tyrosol treatment resulted in 261 and 181 differentially expressed genes with at least a 1.5-fold increase or decrease in expression in *C. parapsilosis*, respectively; however, the initial adherence was not affected by the presence of tyrosol [43]. Interestingly, the ortholog of the *C. albicans CZF1* gene, which is a key transcription factor of biofilm development in *C. parapsilosis*, was upregulated following tyrosol exposure [43,100]. Nevertheless, Jakab et al. (2019) did not observe higher rates of biofilm formation in the presence of tyrosol [43]. In *C. parapsilosis*, tyrosol exposure overexpressed the active efflux pumps and caused an enhanced oxidative stress response, while inhibiting growth, ribosome biogenesis, and virulence. Surprisingly, its metabolism was modulated toward glycolysis and ethanol fermentation [43]. Monteiro et al. (2015) reported that tyrosol exposure did not induce increased adhesion in *C. glabrata* [101].

Regarding tyrosol related toxic effect, initial cytotoxicity was observed at concentrations of >10 mM, 3 mM, 5 mM and >15 mM for human gingival fibroblasts (GN61), human gingival epithelial cells (S-G), human salivary gland carcinoma cells (HSG1) and colon adenocarcinomas cell line (Caco-2), respectively [43,102].

#### *5.2. Antimicrobial Activity of Tyrosol*

Tyrosol is a relatively understudied molecule compared to farnesol in terms of potential antifungal or anti-biofilm activity; despite this, a few studies have examined the potential use of tyrosol in monotherapy or in combination with traditional antifungal agents against *Candida* species [36,72]. Arias et al. (2016) showed that tyrosol treatment at concentrations ranging from 100 to 200 mM exerted a significant reduction in metabolic activity against *C. albicans* and *C. glabrata* two-day-old oral biofilms, which was proportional to a reduction in cell number [103]. Do Vale et al. (2017) showed that tyrosol alone at concentrations of 50 and 90 mM demonstrated inhibition of the planktonic growth of *C. albicans* and *C. glabrata* cells, respectively [104]. However, tyrosol does not significantly reduce metabolic activity or the number of cells for one-day-old oral biofilms; in addition, the nature of interaction of tyrosol with chlorohexidine gluconate was indifferent. Nevertheless, 1.25 mM tyrosol with 0.00725 mM chlorhexidine gluconate showed a synergistic interaction in reducing the number of hyphae formed [104]. A combination of tyrosol and farnesol has been explored for oral *Candida* isolates for both planktonic and sessile growth. This combination showed synergy against *C. glabrata*, indicating that this combination may contribute to the development of oral care products against *Candida* species [105].

In another study, tyrosol showed anti-biofilm activity against denture-derived *C. albicans* isolates. However, it has been shown that the single use of tyrosol cannot decrease hydrolytic enzymes on oral *C. albicans* [106]. Shanmughapriya et al. (2014) observed that tyrosol treatment caused a 25% and a 50% reduction in intrauterine device-derived *Candida krusei* and *C. tropicalis* biofilm production at concentrations of 40 μM and 80 μM, respectively [107]. In addition, amphotericin B combined with tyrosol showed a remarkable inhibitory effect against these non-*albicans* biofilms. A concentration of 4 mg/L amphotericin B in the presence of 80 μM tyrosol exerted approximately 90% inhibition in biofilm formation [107]. Cordeiro et al. (2015) showed that the addition of tyrosol significantly reduced the MICs for amphotericin B, fluconazole, and itraconazole against planktonic *C. albicans* and *C. tropicalis* [108]. Furthermore, exogenous tyrosol alone was able to significantly reduce the biofilm formation of these species at concentrations ranging from 125 to 250 mM. At these concentrations, tyrosol decreased the metabolic activity of growing biofilms by approximately 24 and 30% for *C. albicans* and *C. tropicalis*, respectively. Reduction of metabolic activity was more pronounced when tyrosol was combined with traditional antifungal drugs including amphotericin B, fluconazole, and itraconazole. It should be noted that application of amphotericin B with tyrosol markedly decreased the metabolic activity of mature biofilms (35%) [108]. Kovács et al. (2017) reported that tyrosol may be used as an adjuvant agent with caspofungin or micafungin in alternative treatment strategies [109]. Regarding the in vivo antifungal effect of tyrosol, Jakab et al. (2019) reported that daily treatment with 15 mM tyrosol decreased the fungal tissue burden in their immunocompromised mouse model [43]. In this study, the expression of *ALS6*, which has a pivotal role in adhesion, was significantly reduced by tyrosol treatment. Furthermore, downregulation of the expression of *FAD2* and *FAD3* may also contribute to decreased virulence and kidney fungal burden. The well-documented antifungal effects exerted by tyrosol may be explained by the enhanced oxidative stress and the inhibition of virulence-related genes, growth, and ribosome biogenesis. In addition, tyrosol can alter the metabolism of *Candida* cells toward fermentation [43].

Data on the potential antibacterial effects of tyrosol remain scarce. Arias et al. (2016) found a potential anti-biofilm activity of tyrosol against *S. mutans* in single and mixed species biofilms with *C. albicans* or *C. glabrata* developed on acrylic resin and hydroxyapatite surfaces [103]. Their results may contribute to the development of innovative topical therapies focusing on biofilm-associated oral diseases. Abdel-Rhman et al. (2016) reported substantial antibacterial activity of tyrosol against *S. aureus*; moreover, tyrosol increased susceptibility to gentamicin, amikacin, and ciprofloxacin at subinhibitory concentrations ranging from 3.5 to 14.3 mM [110]. Tyrosol treatment can also influence *S. aureus* virulence, decreasing the production of protease and lipase enzymes and limiting the ability to form biofilms [110]. In the case of *P. aeruginosa*, tyrosol strongly inhibited haemolysin and protease production [111].

#### **6. Future Remarks**

Paradoxically, medical advancement has resulted in an increasing number of immunocompromised individuals susceptible to *Candida* infections. The incidence and mortality rate related to systemic *Candida* infections has remained unchanged, despite the advances in the field of antifungal therapy. Based on recent comprehensive epidemiological studies, the high incidence and mortality may be attributed to sessile *Candida* populations, namely biofilms, which show high resistance against environmental factors, immune responses, and traditional antifungal therapy. Although there is no definitive solution or highly effective therapeutic recommendation against *Candida* biofilms, there are many promising therapeutic strategies including antifungal "lock" therapy, photodynamic inactivation, and the use of natural products or synthetic peptides with antifungal activity. A further solution may be the utilization of quorum-sensing molecules alone or in combination with traditional antifungal agents; however, there are numerous open questions as to their exact action or the interaction between quorum-sensing molecules and the host. In addition, the full understanding of quorum sensing in non-*albicans* species has remained unelucidated. In this review, we provided an overview on the current status of studies focusing on anti-biofilm activity of farnesol and tyrosol. Hopefully, these in vitro and in vivo results can be implemented in therapeutic practice as soon as possible to overcome *Candida* biofilm-related infections.

**Author Contributions:** Conceptualization, methodology and writing were performed by R.K. and L.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** R.K. was supported by the EFOP-3.6.3-VEKOP-16-2017-00009 program and by a FEMS Research and Training Grant (FEMS-GO-2019-502).

**Conflicts of Interest:** L.M. received conference travel grants from Cidara, MSD, Astellas and Pfizer. All other authors report no conflicts of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Biofilm Formed by** *Candida haemulonii* **Species Complex: Structural Analysis and Extracellular Matrix Composition**

#### **Lívia S. Ramos 1, Thaís P. Mello 1, Marta H. Branquinha <sup>1</sup> and André L. S. Santos 1,2,\***


Received: 6 March 2020; Accepted: 1 April 2020; Published: 3 April 2020

**Abstract:** *Candida haemulonii* species complex (*C. haemulonii*, *C. duobushaemulonii*, and *C. haemulonii* var. *vulnera*) has emerged as opportunistic, multidrug-resistant yeasts able to cause fungemia. Previously, we showed that *C. haemulonii* complex formed biofilm on polystyrene. Biofilm is a well-known virulence attribute of *Candida* spp. directly associated with drug resistance. In the present study, the architecture and the main extracellular matrix (ECM) components forming the biofilm over polystyrene were investigated in clinical isolates of the *C. haemulonii* complex. We also evaluated the ability of these fungi to form biofilm on catheters used in medical arena. The results revealed that all fungi formed biofilms on polystyrene after 48 h at 37 ◦C. Microscopic analyses demonstrated a dense network of yeasts forming the biofilm structure, with water channels and ECM. Regarding ECM, proteins and carbohydrates were the main components, followed by nucleic acids and sterols. Mature biofilms were also detected on late bladder (siliconized latex), nasoenteric (polyurethane), and nasogastric (polyvinyl chloride) catheters, with the biomasses being significantly greater than on polystyrene. Collectively, our results demonstrated the ability of the *C. haemulonii* species complex to form biofilm on different types of inert surfaces, which is an incontestable virulence attribute associated with devices-related candidemia in hospitalized individuals.

**Keywords:** *Candida haemulonii* complex; biofilm; extracellular matrix; catheter; polystyrene; virulence

#### **1. Introduction**

*Candida haemulonii*, *Candida duobushaemulonii,* and *Candida haemulonii* var. *vulnera* form a fungal complex (named *C. haemulonii* complex) that is represented by emergent, opportunistic yeasts able to cause human infections with a wide range of clinical manifestations, varying from superficial to deep-seated infections, especially in individuals with immunocompromising health conditions [1]. In this sense, the main isolation sites of the *C. haemulonii* species complex described in the literature are blood, foot ulcers, nails, bones, skin wounds, and vagina; however, there are reports of isolates obtained from other body fluids such as cerebrospinal fluid, bronchoalveolar lavage, vaginal discharge, pleural effusion, peritoneal and ascitic fluids, bile, and urine [1–15].

The multidrug-resistance profile of the *C. haemulonii* species complex has been highlighted by many research groups worldwide, making it a challenge to treat such infections, which is aggravated by the immunological status of the majority of target patients. Although the knowledge about this fungal complex has been growing in recent years, many aspects related to its virulence need to be better investigated. In this sense, biofilm formation is an unquestionable and well-known virulence attribute associated with both bacterial and fungal infections around the world. Biofilm formation by the *C. haemulonii* species complex has already been reported based on the use of classical methodologies [1,7,16], but there is lack of information about the characteristics of the biofilm formed by these fungi. Indeed, it is believed that biofilm lifestyle is the preferred organization mode of microorganisms in nature, which is characterized by a highly complex structured community of microorganisms that interact with each other and with a biotic/abiotic surface, covered by a self-produced extracellular matrix (ECM) composed mainly of proteins, polysaccharides, lipids, nucleic acids, minerals, and water [17,18]. Functionally, the ECM plays an important role in the biofilm maintenance, architecture, and dynamic, being responsible for conferring protection against external stressors, such as host immune responses (both humoral and cellular components) and drugs (either disinfectants or antimicrobial agents), which directly impact the treatment, especially that of seriously ill patients [18,19].

Biofilm-related infections are considered a huge problem in healthcare settings worldwide [20]. Many chronic infections caused by both bacteria and fungi have been associated to biofilm mode of growth, including lung infections (e.g., fungal ball) and chronic leg wounds [20]. *Candida* species, for example, can form biofilm on a variety of medical devices, and it is well-known that catheter-related fungemia is associated with high morbidity and mortality rates among patients in healthcare services, despite the consequent financial burden related to this situation [20]. *C. haemulonii* complex has already been associated to cases of catheter-related fungemia in both pediatric and elderly patients [6,21], and the catheter, in this scenario, acts as a gateway to the infection development as well as to its chronicity.

In the present study, we aimed to investigate the architecture of the biofilm formed by 12 clinical isolates comprising the *C. haemulonii* species complex (*C. haemulonii*, *n* = 5; *C. duobushaemulonii*, *n* = 4; and *C. haemulonii* var. *vulnera*, *n* = 3) on polystyrene, with a special focus on the study of the chemical composition of their ECM. Additionally, we evaluated and compared the ability of these fungal isolates to form biofilm on different medical devices commonly applied in clinical settings, such as nasogastric, late bladder, and nasoenteric catheters made of polyvinyl chloride, siliconized latex, and polyurethane, respectively.

#### **2. Materials and Methods**

#### *2.1. Microorganisms and Growth Conditions*

A total of 12 clinical isolates recovered from patients from Brazilian hospitals between 2005 and 2013 and identified by molecular approaches as belonging to the *C. haemulonii* species complex were used in the present work [10]. Some relevant data about the fungal isolates are summarized in Table 1. Fungal cells were cultured in Sabouraud dextrose medium (under the following conditions: 37 ◦C for 48 h at 200 rpm) and then used in all the experiments. The yeast cells were quantified using a Neubauer chamber.

#### *2.2. Biofilm Formation on Polystyrene*

Fungal cell suspensions in Sabouraud broth (200 μL containing 106 yeasts) were transferred into each well of a flat-bottom 96-well polystyrene microtiter plate, and then incubated without agitation at 37 ◦C for 48 h. Plate wells containing only culture medium were used to set up the reader as blanks. The supernatant fluids were removed by pipetting and, subsequently, the plate wells were washed three times with phosphate-buffered saline (PBS, pH 7.2) to remove nonadherent cells. The measurements of biofilm parameters (biomass, metabolic activity, and ECM) were then performed as described below.


**Table 1.** Clinical isolates used in the present work.

#### *2.3. Biofilm Parameters*

#### 2.3.1. Biomass

Biomass quantification was performed as described by Peeters et al. [22]. Firstly, methanol at 99% (200 μL) was used to fix the biofilms for 15 min at room temperature, then the supernatant was discarded, and the plates were air-dried during 5 min. Afterwards, the plates were incubated for 20 min at room temperature with 0.4% crystal violet solution (200 μL; stock solution diluted in PBS; Sigma-Aldrich, St Louis, MO, USA). The plate wells were finally washed once with PBS in order to remove the excess of staining and the bound dye was then eluted with 33% acetic acid (200 μL) for 5 min. The acetic acid solution (100 μL) was transferred to a new 96-well plate and the absorbance was measured using a microplate reader at 590 nm (SpectraMax M3; Molecular Devices, Sunnyvale, CA, USA).

#### 2.3.2. Metabolic Activity

The metabolic activity of the biofilm was determined using a colorimetric assay able to measure the metabolic reduction of 2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide (XTT; Sigma-Aldrich, St Louis, MO, USA) to a water-soluble brown formazan product [22]. The XTT/menadione solution was prepared by dissolving 2 mg XTT in 10 mL of pre-warmed PBS, which was supplemented with 100 μL of a stock solution of menadione (0.4 mM in acetone). The XTT/menadione solution (200 μL) was added to the plate wells and incubated at 37◦C for 3 h in the dark. Afterwards, 100 μL of supernatant from each well was transferred to a new microplate and the colorimetric changes were quantified using a microplate reader at 492 nm (SpectraMax M3; Molecular Devices, San Jose, CA, USA).

#### 2.3.3. ECM

The biofilm ECM was quantified according to the method described by Choi et al. [23]. Briefly, 0.1% safranin (200 μL; Sigma-Aldrich, St Louis, MO, USA) diluted in PBS was used to stain the nonfixed biofilms, at room temperature for 5 min. Afterwards, the plate wells were washed once with PBS and the bound dye was eluted with 30% acetic acid (200 μL). Supernatants (100 μL) were transferred to a new 96-well plate and absorbance was quantified using a microplate reader at 530 nm (SpectraMax M3; Molecular Devices, San Jose, CA, USA).

#### *2.4. Biofilm Architecture*

#### 2.4.1. Confocal Laser Scanning Microscopy (CLSM)

Biofilms formed on polystyrene surface for 48 h at 37 ◦C were stained with 5 μg/mL of Calcofluor white (Sigma-Aldrich, San Luis, MO, USA) for 1 h at room temperature, protected from the light [24,25]. Subsequently, the biofilms were washed twice with PBS and covered with *n*-propylgallate for observation under a confocal microscope (Leica TCS SP5 with OBS, Berlin, Germany). Three-dimensional reconstitutions of biofilms were obtained by Fiji (ImageJ2, UW-Madison LOCI, Wisconsin, WI, USA) software [26]. The analysis of images was conducted using *z*-series image stacks from spots of each biofilm chosen randomly.

#### 2.4.2. Scanning Electron Microscopy (SEM)

Biofilms formed on polystyrene coverslips (Nalgene, Thermo Fisher Scientific, Waltham, MA, USA), at 37 ◦C for 48 h, were fixed in a solution made of 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, at 4 ◦C overnight. Then, PBS was used to wash the systems, which were post-fixed with 2% osmium tetroxide for 2 h. Dehydration was done in graded concentrations of acetone (25%–100%). The critical point method was used to dry fungal biofilms, which were then mounted on stubs, coated with gold (20–30 nm), and analyzed using a JEOL JSM 6490LV scanning electron microscope [27,28].

#### *2.5. Biofilm ECM Composition*

#### 2.5.1. Extraction of ECM

Biofilms formed on polystyrene for 48 h at 37 ◦C were washed three times with PBS to remove the medium and nonadherent cells. Then, 200 μL of 1.5 M NaCl was added to each well of the microtiter plate and incubated overnight at 4 ◦C [29]. Finally, the well contents were transferred to a clean tube and filtered through a 0.22-μm membrane (Millipore, São Paulo, SP, Brazil).

#### 2.5.2. Chemical Quantification of the Main Biomolecules

The protein concentration was determined by the method described by Lowry et al. [30], using bovine serum albumin (BSA; Sigma-Aldrich, San Luis, MO, USA) as standard. The carbohydrate concentration was determined by the method described by Dubois et al. [31], with some modifications. Briefly, the experiment was carried out using a polystyrene 96-well microplate, in which 50 μL of the extracellular matrix, 150 μL of sulfuric acid, and 30 μL of 80% phenol were added. The standard curve was made with glucose (Sigma-Aldrich, San Luis, MO, USA). The plate was heated in a water bath for 10 min at 90 ◦C, and then incubated at room temperature for 5 min. Finally, the absorbance was measured at 530 nm using a microplate reader (SpectraMax M3; Molecular Devices, San Jose, CA, USA). The nucleic acids present in ECM were extracted with the Gentra® Puregene® Yeast and G+ Bacteria Kit (Qiagen®, Maryland, MD, USA), according to the manufacturer's protocol, and then quantified using a spectrophotometer (Nano-Vue PlusTM; GE Healthcare, Chicago, IL, USA). The sterol concentration was determined using the AmplexTM Red Cholesterol Assay kit (Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions.

#### *2.6. Biofilm Formation on Distinct Catheters Employed in Clinical Settings*

In order to evaluate the ability of *C. haemulonii* species complex to form biofilm on common medical devices, a nasogastric catheter composed by polyvinyl chloride (Medsonda, Arapoti, PR, Brazil), a late bladder catheter composed by siliconized latex (Sisco, São Paulo, SP, Brazil), and a nasoenteric catheter composed by polyurethane (Solumed, Atuba-Pinhais, PR, Brazil) were selected. Autoclaved scissors were used to cut catheters into pieces of approximately 0.30, 0.70, and 0.36 cm2, respectively, and placed on flat-bottom 96-well microplates. Fungal cell suspensions were placed on the catheters in flat-bottom 96-well plates (using polystyrene substratum as control) in Sabouraud medium (10<sup>6</sup> yeasts in 200 μL) at 37 ◦C for 48 h. Blank controls were prepared by adding only culture medium to the catheters. Then, the catheters were washed three times with PBS to remove nonadherent cells and carefully transferred to a new flat-bottom 96-well microplate, and then the biofilm biomass was measured as described earlier.

#### *2.7. Statistics*

All experiments were performed in triplicate, in three independent experimental sets. The results were analyzed statistically by Student's *t*-test (in the comparisons between two groups) and one-way analysis of variance (ANOVA) (in the comparisons between three or more groups). The correlation tests were determined by Pearson's correlation coefficient (*r*). All analyses were performed using the program GraphPad Prism5. In all analyses, *p*-values of 0.05 or less were considered statistically significant.

#### **3. Results and Discussion**

#### *3.1. Biofilm on Polystyrene Surface: Classical Parameters*

It is known that adhesion represents the first step for biofilm formation, which is an important virulence attribute described for several *Candida* species presenting medical implications [32,33]. The relevance of biofilm formation by *Candida* spp. lies the crucial characteristics such as greater resistance to antifungal drugs, host immune responses, and stress situations, resulting in difficulties in the treatment and possible persistence of the infectious process [17]. Taking this into consideration, initially, we confirmed the capacity of clinical isolates belonging to the *C. haemulonii* complex to form biofilm over a polystyrene surface [16]. In this set of experiments, three classical parameters related to biofilm formation were evaluated after 48 h of contact with polystyrene: (i) biomass by the incorporation of crystal violet in methanol-fixed cells, (ii) metabolic activity (cell viability) by reduction of XTT, and (iii) ECM by absorption of safranin, in the latter cases, using non-fixed fungal cells. All 12 clinical isolates comprising the *C. haemulonii* complex formed biofilm at different degrees, exhibiting a typical isolate-specific pattern (Figure 1A,C,E). Statistically significant differences were not observed, while the average measurements of the three biofilm parameters among the three fungal species forming the *C. haemulonii* complex were compared (Figure 1B,D,F). Biofilms revealed by the incorporation of crystal violet and safranin showed the presence of a network formed by yeasts and an exuberant ECM, respectively (data not shown).

Regarding the biofilm biomass, we observed that the average of biofilm formation on polystyrene by the clinical isolates studied herein was similar to that reported by Cendejas-Bueno et al. [1], who also studied clinical isolates of the *C. haemulonii* complex obtained from different isolation sites. The comparison of biofilm formation among the members of other *Candida* species complex has already been documented. In this sense, the three species of the *C. parapsilosis* complex (*C. parapsilosis* sensu strictu, *C. orthopsilosis*, and *C. metapsilosis*) exhibited similar abilities to produce mature biofilms on abiotic surfaces regarding biomass, viability, and three-dimensional architecture [34–36]. Regarding the *C. glabrata* complex, Figueiredo-Carvalho et al. [37] reported that biofilm biomass was significantly higher than *C. nivariensis*.

**Figure 1.** Biofilm formation by the *C. haemulonii* species complex on polystyrene surface. Yeasts (200 μL containing 10<sup>6</sup> cells) were placed to interact with polystyrene for 48 h at 37 ◦C. Afterwards, the systems were processed to detect fungal biomass by crystal violet incorporation in methanol-fixed biofilms at 590 nm, cell viability by the reduction of 2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide (XTT) in formazan at 492 nm, and extracellular matrix by staining non-fixed biofilms with safranin at 530 nm. The results were expressed as absorbance (ABS) values per clinical isolate studied (**A**,**C**,**E**) and mean per fungal species (**B**,**D**,**F**). The results are shown as mean ± standard deviation of three independent experiments. The numbers on the X-axis of graphs represent each of the 12 clinical isolates of the *C. haemulonii* species complex studied, in which *Ch* means *C. haemulonii*, *Cd* means *C. duobushaemulonii,* and *Chv* means *C. haemulonii* var. *vulnera*.

#### *3.2. CLSM Analysis*

The three-dimensional organization as well as the biomass distribution in the biofilms formed by the clinical isolates comprising the *C. haemulonii* complex were analyzed by CLSM (Figure 2), which is a nondestructive technique that allows in situ visualization of the intact biofilm [38]. To do it, Calcofluor white was used to stain the yeasts owing to its affinity to chitin (which is a universal polysaccharide present in the fungal cell wall), evidencing the biofilm biomass as well as the ECM (Figure 2), which is evidenced by the diffuse marking between the yeasts, as previously proposed [39]. In addition, the three-dimensional representation of biofilms was used to determine their thickness, which ranged from 21.6 to 39.1 μm (overall mean = 28.3 ± 5.6 μm) for all clinical isolates studied. The biofilm thickness in each fungal species is as follows: *C. haemulonii*, 21.6 to 32.1 μm (overall mean = 26.1 ± 4.8 μm); *C. duobushaemulonii*, 25.9 to 39.1 μm (mean = 30.5 ± 5.8 μm); and *C. haemulonii* var. *vulnera*, 26.1 to 37.1 μm (mean = 29.1 ± 7.1 μm) (Figure 2). Some authors have documented different thicknesses of biofilms formed by *Candida* species, varying from 11 to 13 μm for *C. tropicalis* [40], 25 to 77 μm for *C. albicans* [39,41], 35.2 to 81.2 μm for *C. famata* [42], and 21 to 26 μm for *C. auris* [43]. In this sense, a variety of conditions can interfere with biofilm features, including isolate specificities, planktonic growth, initial inoculum concentration, and variability on biofilm-forming conditions

(substratum, temperature, CO2 tension, fluid flow, developmental timing, and medium used to support biofilm formation) [44].

**Figure 2.** Representative confocal laser scanning microscopy (CLSM) images of the biofilms formed by the *C. haemulonii* species complex on polystyrene surface. Yeasts (200 μL containing 10<sup>6</sup> cells) were placed to interact with polystyrene for 48 h at 37 ◦C. Subsequently, the biofilms were stained with Calcofluor white, evidencing the fungal biomass. The panels on the left represent the top view images of the fungal biofilms visualized by CLSM; bars represent 5 μm. The graphs on the right represent the three-dimensional reconstruction of the biofilms formed by each species. The isolates *C. haemulonii* (LIP*Ch*4), *C. duobushaemulonii* (LIP*Ch*6), and *C. haemulonii* var*. vulnera* (LIP*Ch*5), which formed the most robust biofilms (**A**), as well as the isolates *C. haemulonii* (LIP*Ch*12), *C. duobushaemulonii* (LIP*Ch*8), and *C. haemulonii* var. *vulnera* (LIP*Ch*11), which formed the weakest biofilms (**B**), are shown.

#### *3.3. SEM Analysis*

SEM analysis was used to assess the biofilm ultrastructure and to evidence peculiar morphological characteristics. Mature biofilms of *C. haemulonii* species complex consisted of a dense network of yeast cells, while structures similar to pseudohyphae were scarcely observed in the majority of the isolates studied (Figure 3). As seen through other approaches, isolate-specific differences were also visualized in biofilm ultrastructure. In this sense, the biofilms formed by *C. haemulonii* isolates LIP*Ch*3 and LIP*Ch*4, for example, exhibited a continuous, intimately packed multilayer structure (Figure 3A–E), while in the remaining fungal isolates, the biofilms were formed by a predominantly discontinuous monolayer with cell aggregates (Figure 3G–I). Water channels could also be visualized (Figure 3J), as well as ECM, as exemplified by isolates of *C. haemulonii* (LIP*Ch*4), *C. duobushaemulonii* (LIP*Ch*6), and *C. haemulonii* var. *vulnera* (LIP*Ch*5) (Figures 4A, 4B and 4C, respectively).

Similarly, Silva et al. [45] demonstrated that *C. glabrata* biofilms are also composed only by yeasts, while *C. parapsilosis* sensu strictu and *C. tropicalis* biofilms characteristics vary depending on the strain used. Those authors observed that some *C. parapsilosis* strains formed biofilms containing both yeast and pseudohypha morphologies, while others presented yeast cells only, and these findings showed no relation with the isolation site of each strain [45]. The majority of *C. tropicalis* isolates displayed only yeast cells, but a small number of isolates showed hyphal formation, especially appearing as long filaments [45]. The biofilm formed by *C. auris*, which is phylogenetically closer to the *C. haemulonii* species complex, is predominantly composed by budding yeast cells and occasionally pseudohyphae [46]. *C. albicans* biofilms, on the other hand, are classically composed by a basal yeast cell polylayer and an upper region formed by hyphal forms [44].

#### *3.4. ECM Composition*

The ECM of biofilms from different *Candida* species exhibits a heterogeneous nature, which has already been thought to be associated to the roles of these components in biofilm architecture and dynamics [47]. The main components of ECM biofilms from *Candida* spp. are proteins, carbohydrates, lipids, and nucleic acids. Several studies have documented the participation of ECM biofilm in adhesion to surfaces, structural maintenance, defense against external aggressors, signaling, and enzymatic issues; however, the enhanced antimicrobial resistance is the most clinically important phenotype of biofilm mode of growth, which is of special concern in hospital settings [25,38,48]. Herein, we investigated the main classic components of the ECM of *Candida* spp. biofilms: proteins, carbohydrates, nucleic acids, and sterols. Among the evaluated components, proteins (mean of 11.61 ± 8.09 μg/mL for *C. haemulonii*, 2.97 ± 1.16 μg/mL for *C. duobushaemulonii,* and 3.88 ± 2.04 μg/mL for *C. haemulonii* var. *vulnera*) were found in greater quantity in the chemically extracted ECM from mature biofilms of all the clinical isolates, followed by carbohydrates (mean of 4.39 ± 2.30 μg/mL for *C. haemulonii*, 3.20 ± 0.74 μg/mL for *C. duobushaemulonii*, and 2.79 ± 1.42 μg/mL for *C. haemulonii* var. *vulnera*); nucleic acids (mean of 0.093 ± 0.074 μg/mL for *C. haemulonii*, 0.026 ± 0.035 μg/mL for *C. duobushaemulonii,* and 0.048 ± 0.082 μg/mL for *C. haemulonii* var. *vulnera*); and, lastly, sterols (mean of 0.023 ± 0.006 μg/mL for *C. haemulonii*, 0.014 ± 0.005 μg/mL for *C. duobushaemulonii,* and 0.007 ± 0.005 μg/mL for *C. haemulonii* var. *vulnera*) (Figure 5). Sterol amounts in *C. haemulonii* isolates were significantly higher when compared with those in *C. haemulonii* var. *vulnera* (*p* < 0.05; one-way ANOVA, Tukey's multiple comparison test) (Figure 5).

**Figure 3.** Representative low-magnification scanning electron microscopy (SEM) images of the biofilms formed by the *C. haemulonii* species complex on polystyrene surface. Yeasts (200 μL containing 10<sup>6</sup> cells) were placed to interact with polystyrene coverslips for 48 h at 37 ◦C, after which the coverslips were visualized using SEM. The images revealed a dense network of yeast cells. In the panel, the images on the left side exhibit different magnifications of the biofilm formed by the isolate LIP*Ch*4 of *C. haemulonii* (**A**–**E**) while on the right side, it is possible to see the biofilms of isolate LIP*Ch*6 of *C. duobushaemulonii* (**F**,**G**) and LIP*Ch*5 of *C. haemulonii* var. *vulnera* (**H**,**I**). Representative water channels are indicated by white arrows in the image of isolate LIP*Ch*3 of *C. haemulonii* (**J**). Note that the white square in (**A**) is the place that was chosen to be amplified and shown in (**B**), and this logic sequence was used in the left side images from (**A**) to (**D**).

**Figure 4.** Representative high-magnification SEM images of the biofilms formed by the *C. haemulonii* species complex on polystyrene surface, focusing on the extracellular matrix (ECM). Yeasts (200 μL containing 10<sup>6</sup> cells) were placed to interact with polystyrene coverslips for 48 h at 37 ◦C, after which the coverslips were visualized using SEM. The ECM of biofilms of *C. haemulonii* LIP*Ch*4 (**A**), *C. duobushaemulonii* LIP*Ch*6 (**B**), and *C. haemulonii* var. *vulnera* LIP*Ch*5 (**C**) are indicated by symbols. The images clearly reveal the presence of an ECM surrounding and holding the yeast cells together (white thin arrows) as well as connecting the yeasts with the polystyrene surface (white thick arrowheads).

Zarnowski et al. [47] described proteins and carbohydrates as the major components of *C. albicans* ECM biofilm, which included 458 distinct protein activities and three polysaccharides of functional importance (α-1,2 branched α-1,6-mannans associated with unbranched β-1,6-glucans forming a mannan-glucan complex, and β-1,3-glucans in a smaller part). Differences regarding non-*albicans Candida* species biofilms ECM composition were reported many years ago. In this sense, Silva et al. [45] documented that *C. parapsilosis* biofilm ECM exhibited high carbohydrate and low protein contents; on the other hand, *C. tropicalis* exhibited high contents of both carbohydrates and proteins, while *C. glabrata* showed low contents of both carbohydrates and proteins.

**Figure 5.** Main biomolecules forming the extracellular matrix (ECM) of the *C. haemulonii* species complex biofilms on polystyrene surface. Yeasts (200 μL containing 106 cells) were placed to interact with polystyrene for 48 h at 37 ◦C. After that, ECM was extracted and carbohydrates, proteins, nucleic acids, and sterols were quantified as detailed in methodology section. The results were expressed as concentration (μg/mL) of each biomolecule per clinical isolate studied (**A**,**C**,**E**,**G**) and mean concentration per fungal species (**B**,**D**,**F**,**H**). The results are shown as mean ± standard deviation of three independent experiments. The symbol (\*) indicates *p*-values < 0.05 (one-way ANOVA, Tukey's multiple comparison test). The numbers on the X-axis of graphs represent each of the 12 clinical isolates of the *C. haemulonii* complex studied, in which *Ch* means *C. haemulonii*, *Cd* means *C. duobushaemulonii,* and *Chv* means *C. haemulonii* var. *vulnera*.

#### *3.5. Biofilm Formation on Medical Devices*

Catheter-related infections are considered a real problem in the medical arena around the world. Candidemia related to catheter use has already been reported in a variety of *Candida* species, including *C. haemulonii* species complex, resulting from the ability of this and other fungal pathogens to adhere to the catheter surface and, consequently, reach the bloodstream mainly of immunocompromised individuals [6]. For this reason, we decided to evaluate the *C. haemulonii* species complex biofilm formation capacity on the surface of different types of catheters currently used in the hospital environment—a latter bladder catheter made of siliconized latex, a nasoenteric catheter made of

polyurethane, and a nasogastric catheter made of polyvinyl chloride. Biofilm formation on these materials was compared to that on polystyrene, a classical substratum used for biofilm analysis (Figure 1). The clinical isolates of the *C. haemulonii* complex were incubated for 48 h at 37 ◦C with the different materials and the biomass was measured by the incorporation of crystal violet. The results were expressed as absorbance (ABS590)/cm2, as the catheters have different dimensions. Our results stressed that the biofilm formation was significantly bigger over the different catheter types when compared with polystyrene regarding all the clinical isolates tested, demonstrating the risk that these clinical isolates would represent in the hospital settings, especially in individuals using nasogastric, nasoenteric, and urinary catheters (Figure 6(aA,aC,aE)). Additionally, biofilm formation on polyurethane and polyvinyl chloride catheters was comparable, with no significant differences between them (Figure 6(aA,aC,aE)). When comparing the mean biofilm formation per species of the *C. haemulonii* complex between the different substrates, we observed that the biofilms formed on polyurethane and polyvinyl chloride catheters were significantly bigger than on polystyrene for both *C. haemulonii* and *C. duobushaemulonii*. Further, biofilms formed on polyvinyl chloride catheters were significantly bigger than on siliconized latex only for *C. haemulonii* (Figure 6(aB,aD)), while for *C. haemulonii* var. *vulnera*, no differences were observed (Figure 6(aF)). Additionally, in relation to the isolation site, cutaneous (fungal isolates LIP*Ch*2, LIP*Ch*3, LIP*Ch*4, and LIP*Ch*7 of *C. haemulonii*; LIP*Ch*1 and LIP*Ch*6 of *C. duobushaemulonii*; and LIP*Ch*5 of *C. haemulonii* var. *vulnera*) versus fluids (fungal isolates LIP*Ch*12 of *C. haemulonii*; LIP*Ch*8 and LIP*Ch*10 of *C. duobushaemulonii*; and LIP*Ch*9 and LIP*Ch*11 of *C. haemulonii* var. *vulnera*) (Table 1), we observed that biofilm formation on the polyurethane (*p* = 0.0427, unpaired Student's*t*-test) and polyvinyl chloride catheters (*p* = 0.0472, unpaired Student's*t*-test) by the isolates from cases of cutaneous candidiasis was significantly higher when compared with isolates obtained from body fluids (Figure 6b). However, for polystyrene and siliconized latex catheters, no statistically significant differences (*p* > 0.05) were observed in this regard (Figure 6b).

Estivill et al. [49], for example, demonstrated the ability of different *Candida* species (*C. albicans*, *C. parapsilosis*, *C. tropicalis*, *C. glabrata*, and *C. krusei*) to form biofilm on different catheter types, and as observed in our clinical isolates, the biofilms formed on the polyurethane and polyvinyl chloride catheters presented very close values for all species studied. Additionally, our group has previously demonstrated the ability of filamentous fungi from *Scedosporium* spp. and *Lomentospora prolificans* to form biofilm on these same catheters [50].

The biofilm formation capacity of *Candida* spp., with a special focus on *C. albicans*, on medical devices has been extensively studied over time. Indeed, the nature of substratum used really influences the biofilm formation. For example, *C. albicans* form better biofilms on soft materials of dentures than on acrylic surfaces [51]. Similarly, *C. albicans* form better biofilms in silicone elastomer and latex surfaces in comparison with polyvinyl chloride and, on the other hand, construct weaker biofilms on polyurethane and silicone [52]. Interestingly, chemical changes made on the surface of medical devices can also interfere in *C. albicans* biofilm formation. For instance, a significant reduction in biomass and metabolic activity of *C. albicans* biofilm was detected when fungal cells were putted to adhere on polyetherurethane covered with 6% of polyethylene oxide [53]. Such differences should be considered, when possible, in the choice of biomaterials to minimize the development of catheter-related *Candida* infections.

**Figure 6.** Biofilm formation on different catheter types by clinical isolates comprising the *C. haemulonii* complex. Fungal cells (200 μL containing 10<sup>6</sup> cells) were placed to interact with polystyrene (PL) and different types of catheters (siliconized latex, SL; polyurethane, PU; and polyvinyl chloride, PC) for 48 h at 37 ◦C. Subsequently, the biofilm biomass was measured by the crystal violet incorporation (590 nm). (**a**) The results were expressed as ABS590/cm2 for each clinical isolate studied (A,C,E) and mean per each catheter type (B,D,F). Values represent the mean ± standard deviation of three independent experiments. The (\*) indicates *p*-values < 0.05 and (\*\*) *p*-values < 0.01 (one-way ANOVA, Tukey's multiple comparison test). The numbers on the X-axis of the graphs represent each of the 12 clinical isolates of the *C. haemulonii* complex studied. (**b**) Comparison of biofilm biomass produced by the clinical isolates on polystyrene and each catheter type regarding the isolation sites (cutaneous and fluids). The 12 isolates were divided into two groups: cutaneous, including nail and skin (*n* = 7); and fluids, including blood, urine, and bronchoalveolar lavage (*n* = 5) (Table 1). (\*) indicates *p*-values < 0.05 (unpaired Student's *t*-test).

#### **4. Conclusions**

Collectively, the present study demonstrated the ability of the *C. haemulonii* species complex to form biofilm on different types of inert substrates, which is an incontestable virulence attribute associated with catheter-related candidemia in hospitalized individuals, representing a serious problem especially when dealing with multidrug-resistant pathogens such as the *C. haemulonii* species complex. Additionally, our results provide new data about *C. haemulonii* species complex biofilm ECM composition.

**Author Contributions:** All authors conceived and designed the experiments. L.S.R. and T.P.M. performed the experiments. All authors analyzed the data. M.H.B. and A.L.S.S. contributed reagents/materials/analysis tools. All authors wrote and revised the paper. All authors contributed to the research and have read and agree to the published version of the manuscript.

**Funding:** This work was supported by grants from Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES - Financial code 001).

**Acknowledgments:** The authors would like to thank Denise Rocha de Souza (UFRJ) for technical assistance in the experiments and Grasiella Matioszek (UFRJ) for confocal analyses.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Unraveling How** *Candida albicans* **Forms Sexual Biofilms**

#### **Austin M. Perry 1,2, Aaron D. Hernday <sup>1</sup> and Clarissa J. Nobile 1,\***


**\*** Correspondence: cnobile@ucmerced.edu; Tel.: +1-209-228-2427

Received: 21 December 2019; Accepted: 13 January 2020; Published: 15 January 2020

**Abstract:** Biofilms, structured and densely packed communities of microbial cells attached to surfaces, are considered to be the natural growth state for a vast majority of microorganisms. The ability to form biofilms is an important virulence factor for most pathogens, including the opportunistic human fungal pathogen *Candida albicans*. *C. albicans* is one of the most prevalent fungal species of the human microbiota that asymptomatically colonizes healthy individuals. However, *C. albicans* can also cause severe and life-threatening infections when host conditions permit (e.g., through alterations in the host immune system, pH, and resident microbiota). Like many other pathogens, this ability to cause infections depends, in part, on the ability to form biofilms. Once formed, *C. albicans* biofilms are often resistant to antifungal agents and the host immune response, and can act as reservoirs to maintain persistent infections as well as to seed new infections in a host. The majority of *C. albicans* clinical isolates are heterozygous (**a**/α) at the mating type-like (*MTL*) locus, which defines *Candida* mating types, and are capable of forming robust biofilms when cultured in vitro. These "conventional" biofilms, formed by *MTL*-heterozygous (**a**/α) cells, have been the primary focus of *C. albicans* biofilm research to date. Recent work in the field, however, has uncovered novel mechanisms through which biofilms are generated by *C. albicans* cells that are homozygous or hemizygous (**a**/**a**, **a**/Δ, α/α, or α/Δ) at the *MTL* locus. In these studies, the addition of pheromones of the opposite mating type can induce the formation of specialized "sexual" biofilms, either through the addition of synthetic peptide pheromones to the culture, or in response to co-culturing of cells of the opposite mating types. Although sexual biofilms are generally less robust than conventional biofilms, they could serve as a protective niche to support genetic exchange between mating-competent cells, and thus may represent an adaptive mechanism to increase population diversity in dynamic environments. Although conventional and sexual biofilms appear functionally distinct, both types of biofilms are structurally similar, containing yeast, pseudohyphal, and hyphal cells surrounded by an extracellular matrix. Despite their structural similarities, conventional and sexual biofilms appear to be governed by distinct transcriptional networks and signaling pathways, suggesting that they may be adapted for, and responsive to, distinct environmental conditions. Here we review sexual biofilms and compare and contrast them to conventional biofilms of *C. albicans*.

**Keywords:** biofilms; *Candida albicans*; sexual biofilms; pheromone-induced biofilms; mating type-like (*MTL*) locus; white cell; opaque cell; phenotypic states; pheromone signaling; biofilm formation; biofilm development

#### **1. Introduction**

Biofilms are communities of microbial cells that are attached to surfaces and encased in a protective substance called the extracellular matrix [1–5]. Biofilms readily form on surfaces that are biotic (e.g., organs, mucosal and epithelial layers, and teeth) and abiotic (e.g., dentures, catheters, and industrial materials) [1–5]. The biofilm growth state provides the microorganisms inside with a sheltered microenvironment that is buffered against fluctuations in the surrounding environment and is protected from predators, environmental stresses, and mechanical forces that microorganisms would normally encounter in the planktonic (or free-living/free-floating) growth state [1–5]. Due to these adaptive benefits, most microorganisms under natural settings have evolved to spend the majority of their existence in the biofilm growth state [1].

Biofilm formation is a key virulence factor for most pathogens, including *Candida albicans,* which is the most commonly encountered human fungal pathogen in clinical settings [3–7]. *C. albicans* causes a wide variety of infections, ranging from benign mucosal (e.g., yeast infections and thrush) to hematogenously disseminated (bloodstream) candidiasis [6,7]. *Candida* infections are notably serious in immunocompromised individuals, such as AIDS patients, patients undergoing chemotherapy, transplantation patients receiving immunosuppression therapy, and patients with implanted medical devices [8–10]. Although research on *C. albicans* has been ongoing for over 70 years, most work has historically focused on *C. albicans* in its planktonic growth state. Over the last 20 years, however, the biofilm growth state of *C. albicans* has become a major area of research focus. *C. albicans* can form biofilms on abiotic surfaces (e.g., dentures, intravenous catheters, and prosthetic devices), as well as biotic surfaces (e.g., mucosal layers in the oral cavity and genitourinary tract) [3–5]. Once established, the cells within a *C. albicans* biofilm are protected from the host immune response, mechanical perturbations, and chemical stresses, allowing *C. albicans* to persist in the host and potentially cause recalcitrant infections [3–5]. More recently, a specialized "sexual" form of *C. albicans* biofilm has been discovered; although structurally similar to "conventional" biofilms, these "sexual" biofilms have many distinct phenotypic characteristics and generate a unique microenvironment that supports *C. albicans* mating [11,12].

Best known as the most common cause of life-threatening fungal infections in hospital settings, *C. albicans* is also a normal commensal in the majority of healthy humans. Remarkably, *C. albicans* can asymptomatically colonize several diverse regions of the body, including the oral cavity, gastrointestinal tract, skin, and genitourinary tract of humans [13–16]. These niches vary dramatically in terms of pH, nutrient sources and availability, and oxygen content [17,18]. This adaptive plasticity is due, in part, to the ability of *C. albicans* to undergo distinct morphological transitions in response to environmental cues; the best characterized examples include the yeast to hyphal cell transition and the transition between two distinct phenotypic cell types, termed "white" and "opaque" [5,18,19]. We begin by reviewing the white-opaque transition as it is intimately intertwined with the formation of sexual biofilms and mating. Next, we review the pheromone signaling and responses that occur in both white and opaque cell types during sexual biofilm formation and mating. Lastly, we compare and contrast conventional and sexual biofilms and consider the mechanisms through which sexual biofilms may aid in the process of mating.

#### **2. The White–Opaque Transition**

White and opaque cell types are heritably maintained for many generations, and reversible switching between the two cell states occurs stochastically under standard laboratory growth conditions [19]. This balance between the white and opaque states is influenced by specific environmental cues that can bias the switch towards one cell type or the other, or even force all of the cells in a population to adopt the white cell phenotype [18–21]. Approximately 16% of the genome is differentially expressed between the white and opaque cell types, resulting in cells with dramatically different phenotypes and functional attributes [18,22–24]. The morphologies of each cell type are also distinct; white cells are spherical and smooth and give rise to white, dome shaped colonies, whereas opaque cells are oblong and pimpled and form flatter and darker colonies [18,19]. Each state displays distinct metabolic preferences, resulting in striking fitness differences under a variety of environmental conditions [25]. White and opaque cells also respond to environmental conditions

in unique ways; for example, opaque cells, but not white cells, can be induced to form filaments in response to nitrogen or phosphate limitation, while white, but not opaque cells, are induced to form filaments in the presence of serum [26]. The two cell types also display distinct responses to alterations in temperature under standard laboratory growth conditions; white cells are stable at 37 ◦C, while opaque cells revert to the white state en masse at 37 ◦C [18]. Opaque cells, however, can be stabilized at 37 ◦C by specific environmental stimuli, such as anaerobic conditions, elevated carbon dioxide levels, *N-*acetylglucosamine, or nutrient limitations [20,21,27–31]. Interestingly, each cell type also interacts with the host immune system in different ways; for example, white cells secrete a macrophage chemoattractant while opaque cells do not, thus increasing the likelihood for opaque cells to escape macrophage engulfment, possibly allowing them to evade this aspect of the host innate immune response [32].

The ability to undergo the white to opaque transition is controlled by the configuration of a discrete region on chromosome 5 known as the Mating Type-Like (*MTL*) locus [33–35]. The *C. albicans MTL* locus can carry two distinct configurations, **a** and α, which consist of genes that specify the **a** and α mating types, respectively [35]. Most *C. albicans* clinical isolates (~97%) are diploid and exist in the *MTL*-heterozygous (**a**/α) configuration, however a few clinical isolates have been found to exist in the *MTL*-homozygous (**a**/**a** or α/α) configuration [33,34]. *MTL*-heterozygous strains express the sex genes *MTLa1* and *MTL*α*2,* the protein products of which form a heterodimer that directly represses the white to opaque transition by binding to the promoter region of *WOR1,* the master regulator of the opaque cell type, and repressing its transcription [33,36]. *MTL*-homozygous strains contain either *MTLa1* or *MTL*α*2*, but not both, and thus *WOR1* expression is derepressed and switching to the opaque state occurs stochastically at a rate of approximately once every 10<sup>4</sup> cell divisions [19,33,34,36]. The white state is considered to be the default cell type, and is often referred to as the "ground state" of the white-opaque switch, since it does not require activation of any known white to opaque transition regulators, while the opaque state is referred to as the "excited state" of the switch because it requires expression of Wor1, which results in activation of many additional regulatory and non-regulatory genes that are specific to the opaque state [22,37].

Although the vast majority (~97%) of *C. albicans* clinical isolates are heterozygous at the *MTL* locus, and were previously presumed to be "locked" in the white cell state [33,34], recent research has shown that the white to opaque switch may be a much more common occurrence in vivo than previously thought. For example, it is now appreciated that natural *MTL*-heterozygous isolates can undergo white to opaque switching in vitro under elevated levels of CO2 and in the presence of *N*-acetylglucosamine, conditions that resemble that of the gastrointestinal tract; however, unlike *MTL*-homozygous opaque cells, these *MTL*-heterozygous opaque cells appear unable to mate [21]. In *MTL*-heterozygous cells, *HBR1*, which encodes a transcription factor that mediates the hemoglobin response, promotes expression of genes carried at the *MTL*α locus and thus indirectly reinforces the **a**1/α2-mediated repression of *WOR1* and ultimately the repression of white to opaque switching [38,39]. Deletion of one copy of *HBR1* in *MTL*-heterozygous cells results in a substantial reduction in *MTL*α*1* and *MTL*α*2* mRNA expression levels and a slight upregulation of *MTLa1* gene expression; the resulting reduction in **a**1/α2 heterodimer levels allows these cells to behave like **a** cells in regards to switching and mating [38,39]. In another example, deletion of *OFR1*, which encodes a protein of unknown function, enables *MTL*-heterozygous white cells to switch to the opaque state and express both **a**- and α-specific pheromones and pheromone receptors, conferring *ofr1* mutants with the unique ability to mate with opaque cells of any *MTL* configuration [40]. In addition, an *MTL*-homozygous clinical isolate strain P94015, which was observed to drift between "white-like" and "opaque-like" cell states, was found to contain a homozygous nonsense mutation in *EFG1*, which encodes a known repressor of the white to opaque transition [41]. Taken together, physiologically relevant environmental cues, or spontaneously arising loss-of-function mutations, could enable naturally occurring strains to undergo white to opaque switching. Lastly, *MTL*-heterozygous cells can become *MTL*-homozygous through loss of heterozygosity on part or all of chromosome 5. This can occur through local gene

conversion, homozygosis of an entire arm of the chromosome, or through spontaneous loss of one copy of chromosome 5 followed by duplication of the remaining homologous chromosome [42,43]. These loss of heterozygosity events have been shown to occur in response to a wide range of environmental conditions, including exposure to antifungal agents, growth in the presence of sorbose, oxidative stress, and temperature stress [42–46].

In addition to *MTL*-heterozygous cells becoming *MTL*-homozygous, *MTL*-homozygous cells can also become *MTL*-heterozygous through the *C. albicans* parasexual life cycle [47]. During parasex, *MTL*-homozygous opaque cells can become *MTL*-heterozygous by mating with *MTL*-homozygous cells of the opposite mating type; this is termed heterothallic mating [47–49]. Interestingly, opaque cells can also mate with opaque cells of the same mating type, termed homothallic mating, providing a means for genetic exchange within unisexual populations and even between clonal progeny of a single parent cell [50]. Generally, the parasexual life cycle requires that *MTL*-heterozygous white cells undergo loss of heterozygosity at the *MTL* locus followed by switching to the opaque cell state [33,51–53]. The resulting *MTL*-homozygous opaque cells secrete sex-specific pheromones that can cause opaque cells of the opposite mating type to extend mating projections towards the highest pheromone concentration gradient [53]. Once two mating projections fuse, the resulting conjugation bridge allows for nuclear fusion and the formation of a tetraploid nucleus [53]. This structure is stable for several cell divisions, thereby producing tetraploid progeny [49,53]. Specific environmental cues can cause the tetraploid cells to reduce their ploidy state via concerted chromosome loss, thereby completing the parasexual life cycle by producing diploid cells [48,49,54,55]. This concerted chromosome loss, however, can often result in aneuploidy, which is hypothesized to allow *C. albicans* to rapidly adapt to variable environments and harsh conditions [49,55]. While asexual reproduction (e.g., through budding) can benefit *C. albicans* populations by preserving well-adapted genotypes, parasex can generate novel allelic combinations to allow for rapid evolution in changing environments [48,54,55], which may contribute to the remarkable ability of *C. albicans* to colonize diverse niches in the body and to its overall success as a commensal and pathogen [49]. Despite these apparent benefits, parasex has thus far been reported to occur at low rates in vivo [27,47]. Given that ~97% of the *C. albicans* population in vivo is thought to be *MTL*-heterozygous [34], the probability that two *MTL*-homozygous white cells of opposite mating types undergo the multiple steps required for mating simultaneously, and within close enough proximity to detect mating pheromone, seems exceedingly low. Recent research, however, is beginning to uncover that homothallic mating occurs more frequent under specific in vitro environmental conditions, such as glucose starvation and oxidative stress, supporting the idea that homothallic mating may be more common than anticipated in vivo [56]. Intriguingly, parasexual mating is hypothesized to occur at high frequencies within sexual biofilms, which are formed by *MTL*-homozygous white cells in response to mating pheromone [11,12]. Like all *C. albicans* biofilms, the multilayer structure of the sexual biofilm is such that its innermost layers are likely to contain lower levels of oxygen and nutrients than the layers closer to its surface, and thus sexual biofilms could be a niche that supports homothallic mating.

Perhaps the most striking difference between the white and opaque cell types is that opaque cells can mate with other opaque cells, but form severely impaired biofilms, while white cells can form robust biofilms, but are unable to mate [11,12,33,47,52,57,58]. Generally, a *C. albicans* biofilm consists of a basal layer of yeast cells with hyphae and pseudohyphae extending away from the substrate to which they are adhered [5,59,60]. In recent years, it has been shown that *MTL*-heterozygous and *MTL*-homozygous white cells form different types of biofilms in response to different stimuli [11,12,57–59]. *MTL*-heterozygous (**a**/α) cells form robust biofilms in response to shear flow forces and various environmental conditions (e.g., temperature, shifts in pH, etc.), and are termed conventional biofilms [5]. Once formed, conventional biofilms are challenging to treat in clinical settings due to their recalcitrance to antifungal agents, mechanical forces, and the host immune response. Alternatively, sexual biofilms formed by *MTL*-homozygous (**a** or α) white cells in response to mating pheromone are less robust than conventional biofilms [11,12], but as discussed above, they may provide an adaptive niche for mating.

#### **3. Pheromone Signaling and Response**

#### *3.1. Mating Pheromones*

The **a** and α pheromones produced by *C. albicans*, encoded by *MFa1* and *MF*α*1* respectively, play essential roles in the processes of heterothallic and homothallic mating [50,61–63]. Opaque α cells constitutively express high levels of *MF*α*1*, producing a trimeric pheromone precursor peptide, whereas white α cells do not [62]. This α-pheromone precursor peptide is thought to be post-translationally modified by the Kex2 protease and Ste13 dipeptidyl aminopeptidase A, to result in two secreted and identical tridecapeptides with the sequence GFRLTNFGYFEPG and one tetradecapeptide with the sequence GFRLTNFGYFEPGK that represent the mature α pheromones; both the tridecapeptide and tetradecapeptide are capable of eliciting mating responses [62–68]. In contrast, **a** cells only weakly express *MFa1* under standard laboratory conditions [61]. However, when exposed to α-pheromone, white and opaque **a** cells highly express both *MFa1* and *MF*α*1* [50,58]. *MFa1* also encodes a precursor peptide which is predicted to be processed similarly to the **a**-pheromone of *Saccharomyces cerevisiae* [61,69,70]. Initial cleavage from the **a** pheromone precursor peptide is thought to occur via the Ste24 and Axl1 proteases [61,69]. The developing peptide is then further processed by the prenyl-group-adding enzymes Ram1 and Ram2, the prenyl-dependent protease Rce1, and the cysteine-carboxy methyltransferase Ste14 [61,69]. The mature **a**-pheromone is a prenylated tetradecapeptide with the sequence AVRSVSTGNCCSTC, and requires Hst6, an ABC transporter, to leave the cell [61,70,71]. Due to the structural simplicity of α-pheromone and the fact that α-pheromone can be more easily chemically synthesized relative to **a**-pheromone, most pheromone signaling experiments in the field are carried out using **a** cells and the addition of chemically synthesized α-pheromone.

Although both *MFa1* and *MF*α*1* are expressed in **a** cells in response to pheromone, α-pheromone is typically degraded by Bar1, an aspartyl protease, via a phenomenon known as "barrier activity" [72]. Barrier activity promotes heterothallic mating in ascomycetes by preventing pheromone hyperstimulation and by allowing for a recovery from cell cycle arrest [72]. It also inhibits the ability of *C. albicans* to undergo auto-pheromone stimulation and thus prevents homothallic mating. Deletion of *BAR1* in *C. albicans* allows for homothallic mating through an autocrine pathway where opaque **a** cells excrete α-pheromone, which then binds to Ste2, the α-pheromone receptor, on the same cell, leading to self-activation for mating [50]. In addition, glucose starvation and oxidative stress enable unisexual populations of opaque **a** cells to undergo homothallic mating despite high *BAR1* expression levels [56], resulting in auto-activated opaque cells that can mate with other opaque cells of the same mating type [50,56]. These findings suggest that certain strain backgrounds as well as specific niches in the human body can override the phenomenon of barrier activity, allowing for unisexual populations to become activated by pheromone [50,56]. This has important consequences for the parasexual lifecycle of *C. albicans* as homothallism allows for same-sex mating to occur within cell mixtures of the same mating types and between certain strains that are incompatible for heterothallic mating [50]. Given that this mechanism results in pheromone stimulation and mating for unisexual populations of opaque cells, a similar scenario could be envisioned within a sexual biofilm. The biofilm environment may even enhance the rate of homothallic mating by sequestering pheromone and possibly protecting pheromone from degradation within the biofilm structure [11,12]. In addition, within a biofilm, recently divided opaque cells would be held in close proximity to each other, increasing both the likelihood of finding a mate nearby and the frequency of mating between progeny of a single opaque cell. Given that *C. albicans* relies on generating aneuploid progeny for genetic diversity, rather than recombination during meiosis, homothallic mating between clones in this capacity could rapidly and efficiently introduce genetic diversity into a population [50,54,55].

#### *3.2. Pheromone-Signaling Pathway Control*

*C. albicans* employs a conserved Mitogen-Activated Protein Kinase (MAPK) signaling pathway to transduce pheromone signals and alter gene expression [73,74]. This pathway begins with the conserved mating type-specific G-protein coupled receptors (GPCRs), Ste2, expressed on **a** cells to recognize α-pheromone, and Ste3, expressed on α cells to recognize **a**-pheromone [73–75]. Activation of either GPCR results in the dissociation of the Gα subunit (Cag1) from the Gβ subunit (Ste4), and the Gγ subunit (Ste18) of a heterotrimeric G-protein [73–76]. The G-protein subunits then activate Cst20, a kinase that activates the downstream MAPK cascade, consisting of Ste11, Hst7, and Cek1/Cek2 [73–75]. All kinases in this pathway, with the exception of Cst20, are held together in close proximity by the scaffolding protein Cst5 [73–75,77]. Cek1 and Cek2 then activate the transcription factor Cph1 in both white and opaque cells, resulting in the differential expression of white and opaque state-specific genes [58,73]. The activities of Cek1 and Cek2 are regulated by Cpp1, a MAP kinase phosphatase [78]. Interestingly, *STE4*, *CST5*, *CEK1,* and *CEK2* are expressed at lower levels in white cells than opaque cells [79], and their repression contributes to the sterility of white cells as white cells engineered to express *STE4*, *CST5,* and *CEK2* (*CEK1* was not tested) at levels similar to opaque cells have been shown to undergo mating at frequencies approaching that of opaque cells [79]. It is also interesting to note that Cek1 (rather than Cek2) appears to play a major role in opaque cell mating; opaque *cek1* mutants mate at much lower frequencies than opaque *cek2* mutants [78]. The precise contributions of Cek1 and Cek2 to the pheromone response in white and opaque cells is complex and an intriguing research area for future study. Nonetheless, we do know that G-protein signaling pathways, such as this one, are highly conserved among fungal pathogens and are involved in controlling several important developmental processes, including mating, filamentation, and virulence [80].

#### *3.3. Di*ff*erences Between the White and Opaque Cell Pheromone Responses*

When opaque cells sense pheromone of the opposite mating type, they become activated for mating via the MAPK signaling pathway (described above). This pheromone stimulation can occur under a variety of different environmental conditions, including planktonic and biofilm conditions [58,61,62,68]. Interestingly, opaque cells have been observed to respond more efficiently to pheromone in media containing alternative carbon sources (e.g., Spider media) [68]. Additionally, the opaque cell pheromone response can be enhanced under a variety of environmental conditions by deletion of *GPA2,* which encodes a G-protein α-subunit that functions at the beginning of the cyclic AMP-protein kinase A (cAMP-PKA) pathway [68]. This finding suggests that mating may occur more frequently within certain (e.g., specific nutrient limiting) host niches and that there is likely crosstalk between the signaling pathways regulating pheromone (i.e., MAPK) and nutrient sensing (i.e., cAMP-PKA) responses.

The opaque cell pheromone response in *C. albicans* is mediated by the transcription factor Cph1, a homolog of the transcription factor Ste12 in *S. cerevisiae* that is activated by a MAPK signaling pathway and controls genes involved in mating [58,70,73,74,81–83]. In opaque cells responding to pheromone, Cph1 initiates a transcriptional response that results in an upregulation of genes involved in filamentation (e.g., *FGR23*), cell fusion (e.g., *FUS1*, *FIG1*), karyogamy (e.g., *KAR4*), MAPK signaling (e.g., *CEK1*/*2*), and adhesion and virulence (e.g., *HWP1*/*2*, *ECE1*, *SAP4*/*5*/*6*, *RBT1*/*4*) [52,53,58,62,68]. Interestingly, although opaque cells generally grow slower than white cells, genes involved in DNA replication and the cell cycle (e.g., *MCM6*, *MCM7*, *PRI2*, and *POL5a*) are specifically repressed in opaque cells responding to pheromone, suggesting that exposure to pheromone further slows progression out of the G1 phase of the cell cycle [52,53,58,62,68,84]. In opaque **a** cells, *STE2* is upregulated, and the α-pheromone receptor Ste2 becomes localized to the tip of growing cellular extensions known as mating projections or conjugation tubes [11,52,53,58,62]; mating projections are phenotypically similar to hyphae, but lack septa [52,53]. Not surprisingly, transcriptional profiling data revealed that opaque cells forming mating projections in response to pheromone upregulate a subset of the genes associated with filamentation and virulence that are upregulated by white cells forming hyphae in response to serum [62,85]. These findings indicate that there is overlap among genes expressed during hyphal

formation and pheromone treatment, but that there are also several genes that are distinctly expressed in each process [62].

Although *C. albicans* white cells are unable to mate, **a** and α white cells still express pheromone receptors and are thought to respond to pheromone in a Cph1-dependent manner [11,58], albeit at a much slower rate than opaque cells [58]. For example, under standard sexual biofilm conditions, the transcriptional response of opaque cells four hours after pheromone exposure is comparable to that of white cells 24 h after pheromone exposure [58]. Interestingly, the pheromone response in white cells appears to occur primarily under sexual biofilm conditions; in fact, much of the response is lost when white cells are subjected to pheromone under planktonic conditions [52,68]. It is also interesting to note that similar to the pheromone response in opaque cells, sexual biofilm formation is highly dependent on nutrient levels [57,68,86], suggestive again of crosstalk between the pheromone response and nutrient sensing signaling pathways. Despite white cells being unable to mate, many genes involved in MAPK signaling and mating are upregulated in white cells responding to pheromone (e.g., *STE2*, *HST6*, *FIG1*, *FUS1*, *KAR4*), which may be an artefact derived from the co-option of Cph1 by white cells for biofilm formation [58]. In addition, many of the adhesion-, biofilm- and other virulence-associated genes upregulated in opaque cells responding to pheromone are similarly upregulated by white cells responding to pheromone in biofilms (e.g., *RBT1*, *HWP1*/*2*, *ECE1*, *PGA23*/*50*, *SAP5*/*6*) [58]. However, unlike opaque cells, white **a** cells do not experience a halt in their cell cycle upon exposure to α-pheromone [11,52]. Overall, in synthetic pheromone-stimulated biofilms, 116 genes are differentially expressed in both white and opaque cells, white cells uniquely differentially express 147 genes, and opaque cells uniquely differentially express 190 genes [58]. Given that Cph1 is believed to mediate both sexual biofilm formation in white cells and mating in opaque cells in response to pheromone, Cph1 may be involved in mediating a core pheromone response involving filamentation and adhesion that can be modified depending on the epigenetic state of the cell [58,73]. Over the course of evolutionary time, it appears that *C. albicans* has rewired aspects of cell–cell communication to be used for host–pathogen interactions, which may provide insight into the unique history of this opportunistic pathogen. Additional work on the regulatory controls of white and opaque cells may improve our understanding of how transcription factors drift to regulate novel functions.

#### **4. Conventional and Sexual Biofilms**

#### *4.1. Properties of Conventional and Sexual Biofilms Compared*

Conventional and sexual biofilms formed by *C. albicans* are both composed of yeast-form, pseudohyphal, and hyphal cells [5,12,60,86]. The *C. albicans* biofilm life cycle typically begins when planktonic yeast-form cells adhere to a substrate in response to specific environmental stimuli [4,5]. These yeast-form cells proliferate, resulting in a dense mat that is tightly anchored to its substrate. Hyphae and pseudohyphae then begin to grow and extend away from the substrate, providing architectural support for the biofilm. As the growing *C. albicans* biofilm matures, the cells within the biofilm produce extracellular matrix material, composed predominantly of proteins, polysaccharides, and DNA that surrounds all of the cells within the biofilm [4,5]. Once a mature biofilm is formed, daughter yeast-form cells disperse from the biofilm and revert to the planktonic growth state or form new biofilms elsewhere [4,5,87]. Although this generalized biofilm life cycle is common across all *C. albicans* biofilms, the configuration of the *MTL* locus and the phenotypic state of the cells play important roles in determining the environmental stimuli that induce biofilm formation as well as certain unique physical characteristics of the biofilms formed. *MTL*-heterozygous white cells form thick and resilient conventional biofilms in response to specific environmental stimuli, such as shear flow rate and host factors, whereas *MTL*-homozygous white cells form thinner and weaker sexual biofilms in response to mating pheromone [11,12,57,86].

Generally, microorganisms that exist in biofilms are protected from environmental stresses relative to microorganisms that exist planktonically [1]. The extracellular matrix surrounding both *C. albicans* conventional and sexual biofilms acts as a physical barrier inhibiting many compounds, such as antimicrobial agents, from penetrating into the deeper layers of the biofilm [3–5]. Mature conventional biofilms, in particular, are highly resilient to most forms of environmental stress, such as treatment with antifungal agents, exposure to mechanical forces, and attack by the host immune system [4,5]. In addition to the physical barrier provided by the matrix, the resilience of conventional biofilms to antifungal agents is also due to the fact that cells within conventional, but not sexual, biofilms upregulate drug efflux pumps (e.g., Cdr1/2, Mdr1), thereby prohibiting antifungal drugs from reaching lethal concentrations within the biofilm [58,60]. Consistent with this finding, sexual biofilms are much more easily permeated by a variety of compounds than conventional biofilms [12,59]. Interestingly, this phenotype can be partially rescued by the overexpression of *BCR1* [12,59], which encodes the biofilm master regulator of several downstream adhesins, suggesting that cell–cell and/or cell–substrate adherence may also contribute to the recalcitrance of conventional biofilms to antimicrobial compounds. Cells within conventional biofilms are also more tightly adhered to each other and their substrates compared to sexual biofilms [12,58,59]. These differences in adherence are likely due to the upregulation of genes involved in adhesion (e.g., *ALS3*) in conventional biofilms, which are less (if at all) upregulated in sexual biofilms [3–5,58,60]. Additional factors contributing to the drug resistance of conventional biofilms include variation in cell membrane sterol composition and the presence of metabolically dormant persister cells, which can display extreme tolerance to most classes of antifungal drugs [3,4,13,88,89]. We note that these two factors have only been studied in conventional biofilms, and thus whether or not they also are present in sexual biofilms is unknown, and an intriguing area of interest for future studies.

If sexual biofilms do not provide the same protective environment as conventional biofilms, why does *C. albicans* bother to form sexual biofilms in the first place? Given that ~97% of the *C. albicans* population in nature is thought to be *MTL*-heterozygous, the chance that two *MTL*-homozygous white cells of opposite mating types will exist in close enough proximity to undergo the complex steps involved to mate is seemingly unlikely [34]. Even if two opaque cells were in close enough proximity to one another, ambient forces would likely disrupt the pheromone concentration gradient before the cells could find one another and fuse. Since sexual biofilms are not nearly as thick or dense as conventional biofilms, these properties could enable opaque cells to extend mating projections through the biofilm towards other opaque cells, while still being sufficiently dense to maintain pheromone gradients and provide some stability against external forces [11,12]. Consistent with the idea that sexual biofilms provide an optimal environment for mating, white **a** cells produce their own pheromone when responding to α-pheromone, which promotes both homothallic and heterothallic mating [90]. In terms of the host response, white cells are preferentially phagocytosed by macrophages as compared to opaque cells and only white cells secrete a leukocyte chemoattractant [32,91]. Thus, white cells may protect mating opaque cells by acting as decoys to sequester infiltrating host cells [32]. Overall, by stabilizing pheromone gradients and providing an optimal environment for opaque cells to undergo mating, sexual biofilms may promote mating in specialized niches of the body that support white-opaque switching (e.g., the skin).

Cell heterogeneity resulting from the various microenvironments present throughout conventional biofilms is also likely to contribute to biofilm resilience [3]. These microenvironment differences lead to specific gene expression changes within cells in discreet environmental niches of the biofilm, resulting in widespread cellular heterogeneity throughout the biofilm architecture [92]. For example, the innermost regions of conventional biofilms are hypoxic and contain less nutrients and more waste products compared to the outermost regions of the biofilm [93]. These unique microenvironments also enable *C. albicans* to coexist and interact with specific microbial species. For example, the hypoxic inner regions of conventional *C. albicans* biofilms support the growth of obligate anaerobic bacteria, such as *Bacteroides fragilis* and *Clostridium perfringens* [3,93]. Although the microenvironments present in sexual biofilms have not been studied to date, because sexual biofilms are much thinner than conventional biofilms [12,59], there are likely to be fewer opportunities for microenvironments to form. Nonetheless, given their phenotypic differences, the microenvironments of conventional and sexual biofilms are certainly distinct.

Interspecies interactions within polymicrobial biofilms between *C. albicans* and other species (mostly bacteria) have only been studied to date within the context of conventional *C. albicans* biofilms. These interactions can be beneficial or antagonistic in nature. A large proportion of research to date has focused on the beneficial interactions between *C. albicans* and *Staphylococcus* species, such as *Staphylococcus aureus*; these two species are often co-isolated from biofilm infections with high mortality rates in clinical settings [94]. Although these two species can form biofilms independently, initial attachment of *C. albicans* cells to surfaces is enhanced when *C. albicans*is co-inoculated with *S. aureus*[95]. In mature polymicrobial biofilms of *S. aureus* and *C. albicans*, *S. aureus* cells can be found adhered to *C. albicans* hyphae and are present throughout the biofilm structure [95–97]. *S. aureus* is, in fact, known to specifically recognize and bind to the adhesin Als3 on the cell surface of *C. albicans* hyphae, and consistent with this, cells of *C. albicans als3* mutants have been found to interact with significantly fewer *S. aureus* cells than wild-type *C. albicans* cells [96]. Interestingly, *ALS3* expression is reduced in sexual biofilms compared to conventional biofilms [58,60], and thus one may hypothesize that *S. aureus* and *C. albicans* are less likely to co-localize in the context of sexual biofilms. Other structural components of *C. albicans* biofilms are also known to play roles in mixed-species interactions. For example, β-glucans present in the extracellular matrix of *C. albicans* biofilms were found to aid methicillin-resistant *S. aureus* (MRSA) strains in surviving vancomycin, one of the few antibiotics effective against MRSA [3,98]. In terms of antagonistic interactions, *Enterococcus faecalis* can secrete EntV, a bacteriocin that inhibits conventional *C. albicans* biofilm formation [3,99]. In another example, *Pseudomonas aeruginosa* can secrete a 12-carbon acyl homoserine lactone that hinders *C. albicans* filamentation and conventional biofilm formation by mimicking farnesol, a quorum sensing molecule produced by *C. albicans* that modulates filamentation [100,101]. *P. aeruginosa* can also release phenazines that specifically inhibit *C. albicans* filamentation and conventional biofilm formation [102]. Overall, given that sexual and conventional biofilms have different physical and biochemical properties, the interactions of these two biofilm systems with other microorganisms are likely to differ considerably.

Conventional and sexual biofilms also differ in their interactions with the host immune response. Neutrophils and mononuclear leukocytes are important host players against fungal infections [103,104]. When neutrophils recognize *C. albicans* cells, they activate a number of antimicrobial defenses, including phagocytosis, degranulation, the release of reactive oxygen species (ROS), and the release of web-like neutrophil extracellular traps (NETs) [103]. In general, neutrophils are very effective at killing planktonic *C. albicans* yeast and hyphal cells [105], where these antimicrobial mechanisms work efficiently. When it comes to *C. albicans* conventional biofilms, however, neutrophils are generally unable to penetrate beyond the outermost regions of the biofilm, ROS are not produced, and NETs are not released [3,4,59,106,107]. This biofilm-specific recalcitrance to neutrophils is largely due to the presence of the extracellular matrix, as physical disruption of the matrix in conventional biofilms restores the ability of neutrophils to release NETs [106]. Consistently, neutrophils are able to release NETs and kill *C. albicans* cells within a biofilm formed by the *C. albicans pmr1* mutant, which is unable to produce matrix mannan [106]. Interestingly, in the presence of a sexual biofilm, neutrophils can penetrate into the innermost layers of the biofilm [59], although whether NETs are released, and fungal cells are killed is unknown. Based on this information, one would hypothesize that sexual biofilms are more susceptible to clearance by neutrophils than conventional biofilms.

In terms of mononuclear leukocytes, these host cells typically respond to *C. albicans* infection by phagocytosing invading cells and releasing cytokines [108]. *C. albicans* cells in conventional biofilms are two to three times more resistant to killing by mononuclear leukocytes than cells growing planktonically [103,108]. In addition, *C. albicans* cells growing in conventional biofilms are capable of altering the cytokine profile of attacking mononuclear cells [108]. For example, the presence of a conventional biofilm leads to the downregulation of TNF-α, a pro-inflammatory cytokine produced by mononuclear leukocytes that would normally suppress biofilm formation [103,108,109]. Intriguingly, conventional biofilms that are grown in the presence of mononuclear leukocytes form thicker biofilms, a phenomenon that is thought to be mediated by an unknown soluble factor that is present when the two are co-cultured [108]. Whether or not this process also occurs with sexual biofilms in the presence of mononuclear cells is unknown, but an interesting area for future exploration.

The host response to *C. albicans* infection is typically initiated by the interaction of host pattern recognition receptors and pathogen-associated molecular patterns (PAMPs) and involves secretion of a variety of antimicrobial compounds. Interestingly, several characteristics of conventional and sexual biofilms inhibit the recognition of PAMPs. For example, hyphal cells, a major component of both conventional and sexual biofilms, are able to 'mask' the β-glucan in their cell walls, blocking a key PAMP recognized by many host immune cell types [4,110,111]. In addition, several cell surface and secreted proteins are capable of sequestering and inactivating host complement proteins, and other secreted anti-immune proteins are expressed at higher levels in conventional biofilms than in planktonic cells [3,4,60]. Although studies to date have only examined conventional biofilms, it seems likely that sexual biofilms would also retain some of these host response characteristics. In fact, we know that some cell surface and secreted proteins involved in inactivating the host immune response (e.g., *SAP4, MSB2*) are also upregulated in sexual biofilms [58]. Nonetheless, how sexual biofilms interact with the immune system and how they compare to conventional biofilms in this regard has not been investigated to date.

#### *4.2. Genetic Regulation of Conventional and Sexual Biofilms*

Our current knowledge of the regulation of conventional and sexual biofilms is summarized in Figure 1. Given that there are many phenotypic differences between conventional and sexual biofilms, it seems likely that the genetic regulation and transcriptional profiles of these two systems should differ as well. As discussed above, the signaling pathway that triggers the formation of sexual biofilms is a MAPK cascade initiated by the pheromone receptors Ste2 or Ste3 [73–75]. This pathway is unique to sexual biofilms, as a Ras1/cAMP pathway that includes Cdc35, Tpk2, and an unknown receptor has been shown to trigger conventional biofilm formation [59,112,113]. In the conventional biofilm pathway, Ras1 activation results in cAMP production, and increased concentrations of cAMP stimulate PKA to initiate the complex transcriptional network controlling conventional biofilm formation [59,112,113]. When comparing the transcriptional profiles of *MTL-*heterozygous white cells grown planktonically versus in conventional biofilm conditions, and white **a** cells grown in sexual biofilm conditions with and without the presence of α-pheromone, there are 662 genes that are induced twofold or more in conventional biofilms, 486 genes that are induced twofold or more in sexual biofilms, and 128 genes similarly induced twofold or more in both systems (examples include *HWP1*, *SAP4*, *SAP5*, *ALS1*) [58,60]. In addition, 187 genes are repressed at least twofold in conventional biofilms, 355 genes are repressed at least twofold in sexual biofilms, and only 19 genes are similarly repressed at least twofold in both systems [58,60]. The dramatic differences in transcriptomic profiles between sexual and conventional biofilms strongly supports the idea that distinct transcriptional networks regulate the formation of these two structures.

**Figure 1.** Summary of the regulation of *C. albicans* conventional (mating type-like (*MTL*)-heterozygous) and sexual (*MTL*-homozygous) biofilm formation and their phenotypic characteristics. Arrows with smaller heads indicate activation (e.g., shear flow and environmental conditions activate the Ras1/cAMP pathway). Arrows with large heads indicate the lifestyle each biofilm type facilitates (e.g., *MTL*-heterozygous biofilms facilitate a pathogenic lifestyle). T-bars indicate inhibitory relationships (e.g., Gal4 and Rfx2 inhibit conventional biofilm formation and conventional biofilms inhibit the deleterious effects of antifungal drugs, mechanical stress and immune attack).

The core transcriptional network controlling conventional biofilm formation consists of nine transcription factors: Tec1, Ndt80, Rob1, Brg1, Bcr1, Efg1, Flo8, Gal4, and Rfx2 [60,114]. By screening a mutant library containing 165 strains with homozygous deletions of genes encoding DNA-binding proteins, a transcriptional network of six transcription factors was identified (Tec1, Ndt80, Rob1, Brg1, Bcr1, Efg1), whose deletion hindered conventional biofilm formation in vitro and in vivo [60]. Interestingly, two of these transcription factor mutants were defective in one in vivo model of biofilm formation but not in another (e.g., the *bcr1* mutant was severely defective in the rat catheter model, but formed a decent biofilm in the rat denture model, while the *brg1* mutant formed normal biofilms in the catheter model, but was severely defective in the denture model) [60]. These findings suggest that the genetic regulation of conventional biofilms may be different depending on the environment [60]. Further investigation into the transcriptional regulators of conventional biofilm formation in a temporal biofilm study revealed three additional core regulators: Flo8, Gal4 and Rfx2 [114]. Interestingly, deletion of *GAL4* and *RFX2* resulted in generally enhanced conventional biofilms relative to wildtype, indicating that they may serve as negative regulators of the network [114]. In order to understand how these transcription factors regulate conventional biofilm formation, genome-wide chromatin immunoprecipitation and microarray experiments were performed on each transcription factor and transcription factor mutant, respectively. These experiments revealed that each of the nine transcription factors contribute to the formation of a complex network that encompasses about 1000 downstream "target" genes [60]. Furthermore, extensive binding between the nine transcription factors and their respective *cis*-regulatory regions highlights a complex set of regulatory feedback loops within the core of the biofilm regulatory network [60,114,115]. Overall, the majority of TFs involved in the conventional biofilm network act as both positive and negative regulators of

various downstream target genes, with the exception of Tec1, which seems to act primarily as an activator [60]. Although the core transcriptional network regulating conventional biofilm formation has been identified, many additional transcription factors have been found to regulate certain aspects of conventional biofilm formation. For example, Rlm1 and Zap1 are both involved in the regulation of the extracellular matrix [116–118]. As we continue research on biofilms into the future, there will certainly be additional regulators identified to play important roles in different aspects of conventional biofilm formation, as well as an increase in our knowledge of the core regulators of sexual biofilm formation.

Sexual biofilms are currently known to rely on four of the nine core transcription factors involved in the conventional biofilm network: Bcr1, Rob1, Brg1, and Tec1 [58]. Deletion of any of these four transcription factors results in a significant decrease in sexual biofilm thickness relative to wildtype [58]. Deletion of *EFG1* does not hinder sexual biofilm formation [58]; rather, the *efg1* mutant appears to form equally thick sexual biofilms relative to wildtype, indicating that *EFG1* is not required for sexual biofilm formation [58]. Interestingly, the *ndt80* mutant forms thicker sexual biofilms than wildtype, although this may not be due to Ndt80 acting as a negative regulator of sexual biofilm formation since deletion of *NDT80* leads to the misregulation of cell separation genes, specifically *SUN41* and *CHT3* [58]. This could consequently result in thicker sexual biofilms as an artifact of enhanced cell clumping and/or reduced cell dispersion during sexual biofilm formation. The fact that this does not occur in conventional biofilms, and that Ndt80 is in fact required for conventional biofilm formation, is an intriguing area for future research. The roles of the other three core transcription factors involved in regulating conventional biofilm formation—Flo8, Gal4 and Rfx2—have not yet been explored in terms of sexual biofilm formation and is another area of interest for future research. Finally, the transcription factor Cph1, which is not required for conventional biofilm formation, plays a central role in the regulation of sexual biofilm formation [58,60]. Deletion of *CPH1* results in the complete obliteration of sexual biofilm formation, and it has been hypothesized that Cph1 is the terminal transcription factor activated by the MAPK cascade in both white and opaque cells responding to pheromone [58]. These ideas have been challenged, where another group found that although the same GPCR (Ste2/3), MAPK cascade (Ste11, Hst7, Cek1/2) and scaffolding protein (Cst5) are used in both white and opaque cell pheromone responses, there are cell type differences in the terminal transcription factors that are activated by pheromone [74]. In opaque cells, their findings suggest that Cph1 is activated for mating, while in white cells, Tec1 is activated for sexual biofilm formation [74,119,120]. The discrepancies between these two findings may be partially explained by differences in growth conditions utilized by the two groups [11,12,58,74,86]. In fact, the different conditions lead to the formation of sexual biofilms with distinct structural features, and one possibility is that different transcription networks may be involved in the two conditions that depend on distinct environmental cues. Given this information, the terminal transcription factor(s) activated by pheromone-stimulated MAPK signaling in white cells remain to be conclusively determined.

The transcriptional network regulating conventional biofilms has been shown to have evolved fairly recently [60]. By determining the master regulators of sexual biofilm formation and its accompanying transcriptional network, we will be able to explore how two seemingly unrelated transcriptional networks and signaling pathways have evolved to interact with one another. If Cph1 is the terminal transcription factor of the pheromone response in white cells, this would indicate that a conserved signaling cascade and its transcriptional regulator evolved to control a novel set of genes during pheromone activation. We can envision two scenarios where this could occur. First, genes associated with biofilm formation may have come under the direct control of Cph1 by the addition of Cph1 recognition sequences to their promoters. Alternatively, one or several regulators of biofilm formation may have come under the control of Cph1 [115]. In the latter scenario, deemed the "regulator-first" model of the evolution of transcription networks [115], Cph1 would have been directly inserted into the older conventional biofilm network, gaining control of several downstream genes associated with biofilm formation, while adding many of the genes that it previously regulated to the network. This model could account for the large size of transcriptional networks (e.g., the conventional biofilm network comprises approximately 20% of the genome), and the reason why complex transcriptional networks include such large numbers of seemingly extraneous target genes [58,60,115]. Since white cells are unable to mate, their main purpose is to form biofilms in response to pheromone, thus reason dictates that they have no need to express genes involved in mating when stimulated by pheromone. Yet, the expression of mating genes has been observed in white cells responding to mating pheromone, where there is a clear induction of genes involved in cell fusion, karyogamy and other aspects of mating (e.g., *FUS1* and *KAR4*) [58]. This regulator-first model is consistent with the hypothesis that Cph1 is the terminal transcription factor activated by the MAPK cascade in both white and opaque cells responding to pheromone. In the alternative hypothesis, Tec1, whose expression is only induced in conventional biofilms via Efg1, may have come under direct control of a novel signaling pathway, namely the pheromone response MAPK cascade. In this scenario, Tec1 would still regulate many of the genes it traditionally regulated and the transcriptional profile of the white cell pheromone response would look similar to conventional biofilm formation. Given that we see a dramatic change in transcriptional profiles between the two biofilm systems and the activation of so many extraneous genes involved in mating in white cells responding to pheromone, we favor the regulator-first model for the evolution of the sexual biofilm transcriptional network.

#### **5. Conclusions**

Sexual biofilms represent a specialized kind of biofilm formed by *MTL-*homozygous cells responding to mating pheromone. The physical characteristics of sexual biofilms differ dramatically from conventional biofilms; indeed, they appear to lack the major characteristics that contribute to the highly pathogenic nature of conventional biofilms. The molecular differences that result in such distinct phenotypes between the two systems remain to be determined. The significance of the unusual characteristics of sexual biofilms and their roles in the lifecycle of *C. albicans* is also not clearly understood. The low frequency of *MTL-*homozygous strains observed in nature and the apparent lack of opaque-specific niches outside of the laboratory led to questions about the existence of a parasexual lifecycle in *C. albicans*in nature. However, it is now appreciated that sexual biofilms may serve as a permeable and penetrable, yet protective, microenvironment that promotes mating in *C. albicans*. Although no in vivo model has been established to investigate the relevance of sexual biofilms in the host, the apparent disadvantageous properties of sexual biofilms for survival in the host may be outweighed by their ability to promote parasexual mating. Future work on the genetic regulation and molecular mechanisms of sexual biofilm formation will improve our understanding of the significance of sexual biofilms as well as the relevance of phenotypic switching and parasexual mating in the lifecycle of *C. albicans* in nature. Overall, the molecular and genetic regulation of conventional and sexual biofilm formation is quite different between the two systems. Conventional biofilms are modulated by the Ras1/cAMP signaling pathway, whereas sexual biofilms are modulated by a MAP kinase pathway; each activating a largely distinct set of transcription factors and likely different transcriptional networks. Understanding how these two transcriptional networks regulate their target genes to give rise to similar yet distinct phenotypes will also provide a basis for studies on the evolution of biofilm formation. Current and future research into sexual biofilms should provide a wealth of knowledge into the molecular genetics, pathogenesis, and evolutionary history of one of the most pervasive fungal pathogens of humans.

**Funding:** This research was funded by the National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) and National Institute of General Medical Sciences (NIGMS) awards R21AI125801 and R35GM124594, respectively, to C.J.N., and by a Pew Biomedical Scholar Award from the Pew Charitable Trusts to C.J.N. This work was also supported by the Kamangar family in the form of an endowed chair to C.J.N. A.D.H. acknowledges funding from the NIH NIAID award R15AI379755.

**Acknowledgments:** The authors thank all members of the Nobile and Hernday labs for insightful discussions on the manuscript as well as the anonymous reviewers for their insightful comments on the manuscript.

**Conflicts of Interest:** Clarissa J. Nobile is a cofounder of BioSynesis, Inc., a company developing inhibitors and diagnostics of biofilm formation. The company has no role in the manuscript. The funders had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Journal of Fungi* Editorial Office E-mail: jof@mdpi.com www.mdpi.com/journal/jof

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-5417-4