Probiotic Yeasts Inhibit Virulence of Non-albicans Candida Species

Non-albicans Candida-associated infections have emerged as a major risk factor in the hospitalized and immunecompromised patients. Besides, antifungal-associated complications occur more frequently with these non-albicans Candida species than with C. albicans. Therefore, as an alternative approach to combat these widespread non-albicans Candida-associated infections, here we showed the probiotic effect of two yeasts, Saccharomyces cerevisiae (strain KTP) and Issatchenkia occidentalis (ApC), in preventing adhesion and biofilm formation of five non-albicans Candida strains, Candida tropicalis, Candida krusei, Candida glabrata, Candida parapsilosis, and Candida auris. The result would influence the current trend of the conversion of conventional antimicrobial therapy into beneficial probiotic microbe-associated antimicrobial treatment.

. For the coinoculation conditions (indicated with a plus sign), probiotic yeasts and non-albicans Candida strains were incubated together for 3 h. For the postinoculation conditions (indicated with a vertical arrow), non-albicans Candida strains were applied to an abiotic surface for 60 min prior to seeding of probiotic isolates and were incubated for an additional 120 min. Crystal violet (0.5%) staining was used to quantify adhered non-albicans Candida cells on abiotic surfaces. as treatment with the commercially available, reference probiotic strain S. boulardii. For all non-albicans Candida biofilms tested, probiotic yeasts inhibited further development of early biofilms compared to the intermediate and mature stages of biofilms ( Fig. 2A  and B). In addition, probiotic strains coincubated with C. krusei, C. glabrata, and C. parapsilosis exhibited 65% to 70% inhibition of biofilm formation whereas C. tropicalis biofilms were inhibited by 44%.
Biofilms, in nature, are found as surface-attached communities of microorganisms. Therefore, we tested the ability of the probiotic yeasts to inhibit biofilms formed by a mixed culture consisting of each of the non-albicans Candida strains in combination with C. albicans. Early (90-min) stages of the mixed-species biofilms showed significant (48% to 81%) reductions upon treatment with putative probiotic yeasts ( Fig. 2D and E). However, intermediate (24-h) and mature (48-h) biofilms were not affected by probiotic treatment even when the duration of treatment was increased (Fig. S3). Furthermore, exposure to probiotics decreased the overall metabolic activity in the biofilm even though the biomass was not affected ( Fig. S3C and D). Together, these results suggest that the putative probiotic yeasts are able to inhibit non-albicans Candida biofilms when applied early in the development.
Putative probiotics reduced filamentation of non-albicans Candida strains. Filamentation is a key virulence factor for various fungi of the Candida species and is positively correlated with adhesion and biofilm formation. Therefore, we wanted to test the effect of the putative probiotics S. cerevisiae and I. occidentalis on cell morphology and filamentation. Levels of hyphal development of C. tropicalis and C. parapsilosis were significantly inhibited upon treatment with probiotic yeasts at 10 8 /ml (Fig. 3). C. krusei and C. glabrata were not tested since they do not exhibit morphological transition. These results bolster the hypothesis that food-derived yeasts function as effective probiotics against pathogenic species of Candida.
S. cerevisiae and I. occidentalis inhibited adhesion of non-albicans Candida strains to cultured epithelial cells from human colon. It is normal to find Candida in small amounts in the mouth, intestines, and skin. Overgrowth of Candida, however, can be problematic. In the context of the human gastrointestinal tract, Candida cells first attach to the epithelial cells and then invade deeper tissues. To test the ability of the putative probiotic yeasts to prevent attachment of Candida to human epithelial cells, we performed cell adhesion assays using monolayers of Caco-2 epithelial cells derived from human colon. The following three conditions were tested: the preinoculation condition, where Caco-2 cells were exposed to probiotics prior to exposure to nonalbicans Candida strains; the coinoculation condition, where Caco-2 cells were exposed to non-albicans Candida and probiotic yeasts simultaneously; and the postinoculation condition, where Caco-2 cells were exposed to non-albicans Candida strains and then treated with probiotic yeasts. Our results indicated that adhesion to Caco-2 monolayer was inhibited by 95% to 99% under the preinoculation condition (data not shown). Under the coinoculation and postinoculation conditions, 72% to 98% of non-albicans Candida strains were inhibited (P Ͻ 0.05) ( Table 1). Interestingly, C. glabrata and C. parapsilosis showed poor adhesion to Caco-2 monolayer compared to C. tropicalis and C. krusei. These results suggest that the probiotic yeasts prevent the attachment of non-albicans Candida yeasts to cultured human epithelial cells.
Application of probiotic yeasts protects C. elegans from non-albicans Candida species. To further investigate the protective phenotypes of the probiotic yeasts in a live animal, we used C. elegans as a model host. C. elegans mimics key aspects of human intestinal physiology, including the presence of polarized microvillus-containing cells (24). The life span of C. elegans reared on a diet of probiotic yeasts (S. cerevisiae, I. occidentalis, or S. boulardii) was similar to that seen with those fed the standard diet of Escherichia coli OP50 (Fig. S4), suggesting that the probiotic treatment is benign and does not affect normal development of the worm.
Coinfection of C. elegans with probiotic yeast along with C. tropicalis, C. krusei, or C. parapsilosis exhibited a life span that was extended (by 5 to 6 days) compared to the life span seen with untreated control worms (Fig. 4). Furthermore, CFU levels of nonalbicans Candida strains recovered from the gut of probiotic-treated worms were significantly decreased compared to the levels seen with the untreated group (Fig. S5). Together, these results suggest that probiotic treatment inhibits the gut colonization of non-albicans Candida strains and extends the nematode life span. In addition, we tested conditions under which the probiotic yeasts were administered after C. elegans worms were infected with C. tropicalis, C. krusei, or C. parapsilosis. The nematodes that were treated with probiotic yeasts postinfection with non-albicans Candida were able to reduce colonization, with no CFU recovered on day 5 after treatment with probiotics. We also noted that worms treated with yeasts of the genus Saccharomyces (both S. cerevisiae and S. boulardii) resisted pathogenic insult better than those treated with the non-Saccharomyces yeast I. occidentalis.
Effect of application of probiotic yeasts on virulence of C. auris. Recently, Candida auris has emerged as a multidrug-resistant superbug that presents a serious global health threat, especially among people with a weakened immune system (25). It attaches to abiotic surfaces, prompting hospitals to take extraordinary decontamination measures, including removal of ceiling and floor tiles, to eradicate it. It has been suggested that the recent rise in C. auris infections is in part due to the overuse of antimicrobial agents that wipe out competing microbes, giving drug-resistant C. auris a chance to overgrow. Therefore, we tested the ability of competing probiotic yeasts to inhibit adhesion of C. auris (obtained from the U.S. Centers for Disease Control). We tested C. auris representing each of the clades. Coinoculation of C. auris strains with probiotics resulted in inhibition of adhesion by 44% to 62% (Fig. 5A), while treatment after probiotics after C. auris had attached to the surface (postinoculation condition) showed a modest decrease in adhesion (34% to 40%; data not shown). We also tested the effect of probiotic treatment on mixed-species biofilms of C. auris and C. albicans, since monocultures of C. auris do not form rich biofilms. Our results revealed that coinoculation of S. cerevisiae or I. occidentalis inhibited mixed-species biofilms of C. auris a For the coinoculation condition (ϩ), probiotic yeasts and non-albicans Candida strains were coinoculated with monolayers of Caco-2 cells and incubated for 3 h. For the postinoculation condition (¡), non-albicans Candida strains were applied on an epithelial layer of Caco-2 cells for 60 min prior to inoculation of probiotic yeasts and incubated for an additional 120 min. Candida chrome agar was used to assess CFU of adhered non-albicans Candida on an epithelial layer of Caco-2 cells. All values are expressed as means Ϯ SD. All the values represent statistical significance at P values of Ͻ0.05 in comparison to the results determined for the untreated control group (-). and C. albicans by 90% (Fig. 5B) and by 27% to 46% (data not shown) under the postinoculation condition. Together, these results indicate that probiotic yeasts inhibit adhesion and biofilm formation of C. auris. The secretome of probiotic yeasts inhibited adhesion of non-albicans Candida to abiotic surfaces. To probe the mechanism of probiotic action, we tested whether the secretome of the probiotic yeasts retained the ability to inhibit adhesion of non-albicans Candida to abiotic surfaces. A two-chamber cell culture insert was used where probiotic yeasts were maintained in the upper chamber and were separated from the lower chamber by a 0.4-m-pore-size membrane that allowed diffusion of small bioactive molecules to the lower chamber containing non-albicans Candida cells (Fig. 6). Our results indicate that a soluble metabolite(s) present in the secretome of probiotic yeasts was able to partially inhibit adhesion of non-albicans Candida (by 22% to 30%) compared to the untreated control. In addition, spent media buffered to neutral pH (pH 7) retained the adhesion-inhibitory effect (18% to 31%), suggesting that antiadhesion nature of probiotic yeast is likely due to a bioactive metabolite(s) and not to the acidic nature of the spent media. These results suggest that a secreted metabolite(s) of probiotic yeasts is able to inhibit the virulence of non-albicans Candida.

DISCUSSION
Biofilm-related clinical complications are a major issue in the health sector. Naturally occurring biofilms are usually polymicrobial, where interaction between microbes may be synergistic or antagonistic. We and others have observed that non-albicans Candida species such as Candida parapsilosis, C. pseudotropicalis, and C. glabrata produced immature biofilms when grown as monocultures compared to C. albicans (26). Fungi of the Candida species have been shown to synergize with bacteria in the oral cavity for adhesion and biofilm formation (27). Likewise, C. glabrata relies on C. albicans for initial adhesion and biofilm development (28). On the other hand, Lactobacillus species have been shown to inhibit the initial stages of biofilm in mixed cultures (29). We used several in vitro and ex vivo tools, including analyses of adhesion to abiotic surfaces and cultured epithelial cells derived from human colon (Caco-2 cell), morphological transition, and biofilm formation, to assess effects of our potential probiotic strains against non-albicans Candida stains. Our results converge on the notion that probiotic yeasts inhibit adhesion and decrease metabolic activity of biofilms, thereby controlling growth and colonization of pathogenic Candida.
We used the nematode Caenorhabditis elegans as a live host model (30) since facets of its innate immune system are faithfully conserved in humans (31). We demonstrated that nematodes treated with probiotic yeasts are better able to withstand pathogenic insult from several non-albicans Candida species. Similarly, Lactobacillus acidophilus was previously shown to significantly decrease the colonization of infectious Gram-positive bacteria in C. elegans gut and to enhance the life span of the worm (32).
Probiotic action can be attributed to physical and/or chemical factors. Here, we provide evidence of the chemical nature of probiotic action since cell-free secretomes of probiotic yeasts retained inhibitory activity. We propose that a secondary metabolite(s) produced by the probiotic yeasts is secreted into the milieu, where it interferes with the pathogenic program of the non-albicans Candida species. Other reports have demonstrated that short-chain fatty acids or bacteriocins showed significant anti- Candida activity for various Candida species (19,(33)(34)(35). Probiotics may also pose a physical barrier by binding surface proteins that promote pathogen attachment or compete for limited nutrients. We showed that metabolically inactive probiotic yeasts retained minimal inhibitory effect, suggesting that the live probiotic cells and their metabolites likely act synergistically. Prior studies using various in vitro and in vivo models have also revealed synergistic mechanisms that involve immune simulation and competitive binding (21,(36)(37)(38).
Small molecules (such as filastatin, farnesoic acid, and gymnemic acids) and hydrophilic or antibody-coated compounds have been proposed as biotherapeutic agents but are associated with significant safety concerns (39)(40)(41)(42). Therefore, probiotics such as S. cerevisiae and I. occidentalis have the potential to inhibit key virulence traits of the most common non-albicans Candida species, i.e., C. tropicalis, C. krusei, C. glabrata, and C. parapsilosis. To meet the growing need for treatment options for biofilm-associated clinical complications, these food-derived yeasts represent a safe and attractive alternative to conventional treatment for Candida infections.
Saccharomyces cerevisiae (strain KTP; accession number MH142729) and Issatchenkia occidentalis (ApC; accession number KF551991) were isolated from toddy and fermented apple juice, respectively. The commercially available Saccharomyces cerevisiae var. boulardii (NCDC363) strain obtained from National Collection Centre for Dairy Cultures, India, was used as a reference strain in the study. The non-albicans Candida species C. tropicalis (MYA 3404), Candida krusei (44), C. glabrata (44), and C. parapsilosis (CDC317) were used in the study. The C. albicans strain (SC5314) was used for mixed-culture biofilm studies with non-albicans Candida stains. All strains were cultured in yeast extract-peptonedextrose (YPD) media at 30°C for 24 h. RPMI 1640 containing L-glutamine, phenol red, 0.2% glucose, and 0.165 M MOPS (morpholinepropanesulfonic acid) buffer without sodium bicarbonate was used to test for plastic adhesion and biofilm formation.