Prostaglandin E2 from Candida albicans Stimulates the Growth of Staphylococcus aureus in Mixed Biofilms

Background Previous studies showed that Staphylococcus aureus and Candida albicans interact synergistically in dual species biofilms resulting in enhanced mortality in animal models. Methodology/Principal Findings The aim of the current study was to test possible candidate molecules which might mediate this synergistic interaction in an in vitro model of mixed biofilms, such as farnesol, tyrosol and prostaglandin (PG) E2. In mono-microbial and dual biofilms of C.albicans wild type strains PGE2 levels between 25 and 250 pg/mL were measured. Similar concentrations of purified PGE2 significantly enhanced S.aureus biofilm formation in a mode comparable to that observed in dual species biofilms. Supernatants of the null mutant deficient in PGE2 production did not stimulate the proliferation of S.aureus and the addition of the cyclooxygenase inhibitor indomethacin blocked the S.aureus biofilm formation in a dose-dependent manner. Additionally, S. aureus biofilm formation was boosted by low and inhibited by high farnesol concentrations. Supernatants of the farnesol-deficient C. albicans ATCC10231 strain significantly enhanced the biofilm formation of S. aureus but at a lower level than the farnesol producer SC5314. However, C. albicans ATCC10231 also produced PGE2 but amounts were significantly lower compared to SC5314. Conclusion/Significance In conclision, we identified C. albicans PGE2 as a key molecule stimulating the growth and biofilm formation of S. aureus in dual S. aureus/C. albicans biofilms, although C. albicans derived farnesol, but not tyrosol, may also contribute to this effect but to a lesser extent.


Introduction
The first study suggesting a specific interaction between Staphylococcus (S.) aureus and Candida (C.) albicans was published in 1976 [1]. Since then, a number of studies corroborated this result and showed a synergistic interaction of S. aureus and C. albicans with enhanced mortality in animal models [2][3][4].
In vitro studies of complex microbial communities show that intra-species and inter-species interactions are mediated via small molecules released into the extracellular environment, i.e. quorum sensing molecules, extracellular virulence factors, or secondary metabolites [5][6][7].
Candida albicans is a commensal microorganism in healthy individuals but is capable of causing disseminated or chronic infections when the host mucosal barrier is breached and the immune response is inadequate. In vivo, inflammation is mediated by the production of eicosanoids including prostaglandins and leukotrienes. Prostaglandin E 2 (PGE 2 ), an oxygenated metabolite of arachidonic acid, is known to regulate the activation, maturation, cytokine release and migration of the mammalian cells, particularly those involved in innate immunity [14][15][16][17][18]. Interestingly, C. albicans produces authentic PGE 2 from external arachidonic acid [19], which is upregulated during biofilm formation suggesting that PGE 2 may represent a significant virulence factor in biofilm-associated infections.
In order to evaluate the possible effect of PGE 2 on bacterial biofilms, we established a S. aureus/C. albicans dual species biofilm model. Using this model we were able to replace the stimulatory activity of C. albicans on S. aureus by synthetic purified PGE 2 , suggesting that this metabolite plays an important role in the interaction of S. aureus and C. albicans in biofilms.
Clinical isolates were cultured from routine diagnostic samples which were sent to the Institute of Medical Microbiology. The samples were processed in the diagnostic laboratory under the special guidance of the author as described below. After homogenization with sterile saline (1:1) the sputum samples were plated on various agar plates and incubated at 37°C and at 30°C for two days, respectively, and at 25°C for further three days. The throat swabs were processed in a similar manner without homogenization. After visible growth the colonies were identified by the use of mass spectrometry (Vitek2 MS; Biomerieux, France) according to the instructions of the manufacturer. All strains identified from patients with cystic fibrosis were routinely stored at -70°C using the Microbank system as recommended by the manufacturer (Bestbion; Colonia, Germany).
For the use in this study, clinical strains were cultured from the Microbank system on Columbia agar plates containing 5% sheep blood (Beckton Dickinson) and incubated over night at 37°C. Subcultures on blood agar plates were incubated for further 24h under the same conditions. From these subcultures glycerin stocks (99.9%; Calbiochem, Heidelberg, Germany) were prepared and stored at -70°C until further use. Dual S. aureus/C. albicans biofilms Three to five S. aureus colonies grown from glycerin stocks were inoculated in 5mL trypcase soy broth (TSB-T; Biomerieux, Nürtingen, Germany) and incubated over night at 37°C with agitation of 180 rpm. C. albicans cells from glycerin stocks were propagated overnight in 100mL Yeast peptone dextrose (YPD) medium (Sigma, Hannover; Germany) at 30°C on an orbital shaker at 180 rpm. Under these conditions all strains grow in the budding yeast phase. Candida cells were harvested by centrifugation, washed twice in sterile phosphate buffered saline (1xPBS; Sigma), re-suspended and adjusted to a cellular density equivalent to 3x10 6 cells/mL. S. aureus cells were adjusted to an optical density of 0.02 in RPMI1640 using a spectrophotometer (Eppendorf, Hamburg, Germany) which is equivalent to approximately 3x10 6 cells/mL [24]. One mL of each microbial suspension was added to each well of a 12-well plate (Greiner, Frickenhausen, Germany). Biofilms were incubated in RPMI1640 buffered with HEPES and supplemented with Lglutamine and 10% heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific, Schwerte, Germany) for the appropriate time at 37°C with 75 rpm of agitation in humid atmosphere. Medium was changed daily. Under these conditions S. aureus and C. albicans formed stable and reproducible mono-microbial and mixed biofilms which was tested using one-way ANOVA (analysis of variance) as described elsewhere [25,26]. After the appropriate time of incubation supernatants were removed. Biofilms were scraped from the bottom of the wells and homogenized with Sputasol liquid (50μg/mL; Thermo Fisher) as described by Efthimiadis et al. [27]. Cells were collected by centrifugation (1100xg for 5 min) and washed twice in sterile PBS. The inoculum was confirmed by quantitative culture (serial diluted 10-fold) on mannitol salt agar (MSA2 agar; Biomerieux) after 24h of incubation at 37°C for S. aureus or on Chromagar CANDIDA (BD) after 48h at 30°C for C. albicans.

Crystal violet staining
Mono-microbial and mixed biofilms were stained with crystal violet (CV) solution as described by Peeters et al. [28]. Briefly, biofilms cultured in 96-well plates were fixed with methanol. Then, 0.2% crystal violet solution was added to each well and incubated for 20 min at room temperature. Excess CV was removed by washing under running tap water and bounded CV was released by 33% acetic acid (Sigma). Absorbance was measured at 570 nm.

Determination of the prostaglandin concentration
Prostaglandin E 2 (PGE 2 ) production was measured from supernatants of different monomicrobial and dual biofilms using a monoclonal PGE 2 enzyme-linked immunosorbent assay (ELISA; Cayman Chemicals, Ann Arbor, USA) according to the instructions of the manufacturer. Supernatants were collected from the biofilms after the appropriate time point, centrifuged at 8000xg, filtered twice by using a 0.22 μm syringe filter (Roth, Karlsruhe, Germany) and stored at -80°C. To confirm, that the supernatants yielded no living cells of S. aureus and C. albicans, 100 μl of the supernatants were plated onto blood agar plates and incubated for 48h at 37°C and for further 48h at 30°C. All supernatants processed by this method were cellfree. In serum, PGE 2 is rapidly degraded into unstable intermediates [19], i.e. 15-keto-13, 14-dihydro-PGE 2 , which were determined with a PGE 2 metabolite kit (Cayman Chemicals).
The PGE 2 metabolite kit converts the unstable intermediates into stable measurable derivative serving as marker for the PGE 2 production. Background levels of PGE 2 detected in RPMI 1640 with and without FBS were subtracted from experimental samples.

Purified substances
Purified tyrosol, farnesol and prostaglandin E 2 (PGE 2 ) were purchased from Sigma-Aldrich. Stock solutions of tyrosol (2-4-hydroxyphenylethanol) and farnesol were dissolved in DMSO. PGE 2 stock solution (1mg/mL) was prepared in ethanol and further diluted in 1xPBS. Established 24h old S. aureus biofilms were incubated overnight in the presence of the appropriate drugs, biofilms were harvested and the numbers of colony forming units (cfu`s) were determined as described above. Control experiments revealed that DMSO and ethanol did not affect to the growth of S. aureus and C. albicans.

Inhibitors
The cyclooxygenase inhibitor indomethacin (Sigma) was dissolved in dimethyl sulfoxide (stock solution 1mM) and further diluted in PBS prior to addition to mixed biofilms. Dual biofilms without supplements or with DMSO alone were used as control. DMSO did not influence the growth of the biofilms.

Statistical analysis
Data represent the mean plus the standard deviation of two independent experiments with three intra-assay replicates.
Statistical significance was determined using Student's t-test with SigmaStat statistical software (version 2.0). P-values 0.05 were considered statistically significant.

Ethics statement
According to the written decision of the clinical research ethics committee of the Otto-von-Guericke University of Magdeburg the current study did not require approval by the local ethics committee because no human material or data attributable to individual patients were used.

Stimulation of S. aureus by C. albicans in mixed biofilms
Initial mixed biofilm experiments with clinical strains of S. aureus and C. albicans showed that biofilm thickness of mixed biofilms was significantly increased compared with S. aureus mono-microbial biofilms (Fig 1A).
Mixed biofilms yielded a significantly higher number of S. aureus colony forming units (cfu`s) compared to mono-microbial S. aureus biofilms with a dramatically increase after 72h of co-culture ( Fig 1B).
In order to investigate, if the growth-stimulating effect of C. albicans depends on the direct contact of bacteria to the hyphal network, 24h old S. aureus biofilms were incubated with cellfree supernatants from C. albicans 31883 and SC5314 diluted 1:1 with fresh medium (Fig 2A). The number of cfu`s of S. aureus increased 10-fold when supernatants derived from 72h or 96h old C. albicans biofilms were added. The strongest increase was observed with supernatants from C. albicans SC5314. Remarkably, the stimulatory activity of supernatants was heatlabile. Incubation of S. aureus with heat-treated supernatants from C. albicans mono-microbial biofilms did not promote the propagation of S. aureus (Fig 2B).

Subinhibitory concentrations of farnesol promote biofilm formation of S. aureus
In order to further define the growth-promoting activity of C. albicans culture supernatants the effect of a number of molecules secreted by C. albicans such as farnesol, tyrosol and prostaglandin E 2 (PGE 2 ), were tested. Synthetic farnesol at concentrations between 5 and 0.5 nM (Fig 3A), but not purified tyrosol, stimulated the growth of S. aureus in biofilms, whereas concentrations 0.5 μM significantly reduced the growth of S. aureus (S1 Fig). The application of supernatants from cultures of both C. albicans SC5314 (farnesol producer) and ATCC10231 (farnesol nonproducer) significantly enhanced the growth rates of S. aureus (Fig 3B). Interestingly, the farnesol-producing strain exhibited a stronger impact to the growth and biofilm formation of S. aureus than the farnesol-negative isolate.
Purified PGE 2 stimulates growth of S. aureus in biofilms Next step was to analyze the effect of purified PGE 2 on 24h old S. aureus biofilms. As shown in Fig 4A, purified PGE 2 significantly enhanced S. aureus biofilm formation in a dose-dependent manner. The stimulatory activity of purified PGE 2 was neutralized by heating treatment of the PGE 2 solution (Fig 4B).
C. albicans PGE 2 -deficient mutant strain do not stimulate growth of S. aureus PGE 2 and its metabolites were abundantly detected in culture supernatants of C. albicans wt and reference strains, and their amount increased according to the C. albicans cell density.
In contrast, PGE 2 levels of the C. albicans null mutant (ura3-/-fet31-/-), deficient in the multi-copper oxidase genes fet31, were below the detection limit of the assay (Fig 5A and S2  Fig).
The concentration of purified PGE 2 inducing the growth of S. aureus were similar to the PGE 2 levels measured in supernatants of C. albicans mono-microbial biofilms (Figs 4A and  5A).
To test the hypothesis that PGE 2 is responsible for the stimulatory effect of C. albicans on S. aureus biofilms, the effect of cell-free supernatants derived from PGE 2 producers (C. albicans SC5314 and C. albicans ura3-/-) and from the PGE 2 nonproducer (C. albicans ura3-/-fet31-/-) to the growth of S. aureus was investigated (Fig 5B). Supernatants derived from the PGE 2 -deficient mutant were unable to stimulate the growth of S. aureus.

Treatment of mixed biofilms with indomethacin reduced growth of S. aureus
In addition, we tested whether the nonspecific cyclooxygenase inhibitor indomethacin can influences the stimulatory activity of C. albicans to S. aureus. Indomethacin at a concentration between 10 and 1000 pg/mL efficiently blocked the biosynthesis of PGE 2 ( Fig 6A) and its metabolites (Fig 6B) by C. albicans. In cultures of mixed S. aureus/C. albicans biofilms indomethacin significantly suppressed the growth of S. aureus in a dose-dependent manner ( Fig  6C).
Experiments performed with the S. aureus small colony variant 31338 revealed similar results.
Neither DMSO nor Indomethacin significantly affected the growth of S. aureus or C. albicans. Data are exemplarily represented for S. aureus 19552 and C. albicans SC5314 (S3 Fig).

S. aureus does not stimulate the PGE 2 synthesis in C. albicans
The levels of PGE 2 and of the PGE 2 metabolites were not enhanced in dual biofilms of S. aureus and C. albicans wild type strains compared with C. albicans mono-microbial biofilms, indicating that S. aureus did not stimulate the release of PGE 2 from C. albicans isolates (S4 Fig).

Discussion
In polymicrobial infections, biofilm formation is the predominant mode of life [29][30][31]. Biofilms are structured microbial communities attached to biotic or abiotic surfaces and embedded in a matrix of exopolymers to withstand the host immune response and exhibit increased resistance to antimicrobial agents [32][33][34][35]. It is estimated that 27% of nosocomial C. albicans bloodstream infections are polymicrobial, with S. aureus as the third most common organism isolated in conjunction with C. albicans [36].
In vitro studies demonstrated that the formation of S. aureus biofilms on abiotic materials in the presence of serum requires pre-coating and nutrient supplementation [37]. In order to study the interactions between C. albicans and S. aureus in mono-microbial and dual species biofilms, we have optimized an in vitro biofilm model of infection. Using this model, we observed increased growth rates of S. aureus in dual biofilms in the presence of C. albicans. In vitro studies described the formation of S. aureus microcolonies on the surface of biofilms with C. albicans serving as the underlying scaffolding [38,39]. The adherence of S. aureus on fungal hyphae was mediated by the Candida agglutinin-like protein 3 (Als3p; [40]). Studies of the protein expression during the growth of dual C. albicans/S. aureus biofilms identified 27 differentially regulated proteins, mostly involved in growth, metabolism, or stress response. Among the down-regulated staphylococcal proteins was the global transcriptional repressor of virulence factors, CodY, suggesting that the enhanced pathogenesis of S. aureus may not only be due to physical interactions [41,42].
Staib et al. [1] hypothesized the potential existence of C. albicans end-products as a substratum for a favorable growth of S. aureus. One of the C. albicans products continuously excreted in the environment is farnesol a quorum sensing molecule involved in the regulation of the fungal yeast-mycelium dimorphism. At high levels, farnesol has been shown to exhibit antimicrobial activity against various pathogens [12]. Kaneko et al. reported growth inhibition of S. aureus at concentrations of 40 μg mL -1 (equivalent to 180 μM) which is concordant with our results [43]. In this study, subinhibitory concentrations of farnesol (0.5-5 nM) promoted the biofilm formation of S. aureus, an effect already described for several antimicrobial agents [44][45][46][47]. In stationary phase cultures of C. albicans, farnesol accumulated at levels between 2 and 4 μM [8,48]. In addition, our data showed that S. aureus biofilm production was propagated by the farnesol-deficient C. albicans strain ATCC 10231 albeit at a lower level compared to the farnesol producer C. albicans SC5314, whereas the PGE 2 null mutant did not enhanced the growth of S. aureus. Unlike SC5314, ATCC10231 excreted significantly lower levels of PGE 2 in this experimental setting indicating that the differences in the growth-stimulating effect to S. aureus resulted from the PGE 2 levels. In conclusion, it appears that PGE 2 has a superior role in the induction of S. aureus biofilm formation as compared to farnesol, which, however, also supports the growth of S. aureus at low concentrations.
We demonstrated for the first time that C.albicans PGE 2 may be one of the suggested endproducts presumed by Staib [1]. As shown in this study, the stimulatory effect to the growth of S. aureus was not observed, when C. albicans supernatants or medium supplemented with purified PGE 2 were heat treated suggesting that PGE 2 is heat-sensitive. This observation is concordant with the results described by Carlson who reported increased mortality rates of mice co-infected with sublethal doses of C. albicans and S. aureus [3]. This effect could not be reproduced when either of the agents was heat inactivated.
Endogenous PGE 2 maintains the integrity of the mucosal barrier in the lungs and in the gastrointestinal tract [49][50][51]. Mucus accumulation and overexpression of inflammatory response genes are relevant pathogenic features of cystic fibrosis (CF), a genetic disorder caused by mutations in the cystic fibrosis transmembrane conductance regulator gene cftr [52]. Recently, Harmon et al. demonstrated a defective peroxisome proliferator-activated receptor (PPARy) function in epithelial cells which was caused by the decreased conversion of PGE 2 to the PPARy ligand 15-keto-PGE 2 leading to the accumulation of PGE 2 [51]. As elevated levels of PGE 2 have been detected in respiratory tract samples of patients suffering from cystic fibrosis [53,54] and in consideration of the results from this study, it seems to be possible that enhanced PGE 2 levels may contribute to the early airway colonization due to S. aureus in patients with CF. PGE 2 was demonstrated to induce germ tube formation and to be involved in biofilm formation by C. albicans [55,56] suggesting C. albicans PGE 2 to be a potential virulence factor. Despite induced endocytosis by host cells at the early stage of infection, hyphal mediated active penetration of the host tissue is the major route of invasion [57,58]. Tissue invasion by C.albicans is associated with cytokine secretion, inducing host immune response mechanisms. Phagocytosis by macrophages and neutrophils represents the first line of defense against Candida infections but intracellular killing is not always effective. One of the underlying mechanisms is the inhibition of the phagosome maturation and the nitric oxid (NO) production in macrophages after infection with C. albicans [59], an effect that may be resulted from candidal PGE 2 excretion. PGE 2 is a critical molecule that regulates the activation, maturation, migration, and cytokine secretion of several immune cells, particular those of the innate immunity. In the context of infection, endogenous PGE 2 inhibits the cytolytic effector function of natural killer (NK) cells, the activation, migration and production of proteolytic enzymes in granulocytes and limits the phagocytosis and pathogen-killing function of alveolar macrophages (reviewed by [14,18]. Aronoff et al. reported the negative regulatory role of endogenously produced and exogenously added PGE 2 on FcRy-mediated phagocytosis of bacterial pathogens by alveolar macrophages suggesting that PGE 2 derived from C. albicans may impair the local host innate immunity [60]. This suggestion is underlined by the investigations of Roux et al. who had shown that airway colonization with C. albicans inhibited phagocytosis of S. aureus and P. aeruginosa and enhanced the prevalence of bacterial pneumonia, which was reduced by antifungal treatment [61,62]. Mice inoculated with either S. aureus or C. albicans survived infection, whereas combined infection with both pathogens enhanced the mortality rate to 40-100% [3,4]. Reversely, enhanced microbial clearance and survival was demonstrated in studies with COX-2-deficient mice [63][64][65]. This is in line with our observation that a mutant strain of C. albicans deficient in PGE 2 production did not promote the growth of S. aureus. Furthermore, in this study the non-selective cyclooxygenase (COX) inhibitor indomethacin that blocked PGE 2 biosynthesis by C. albicans also reduced the growth of S. aureus in dual biofilms to a level observed in mono-microbial S. aureus biofilms.
Finally, in this study S. aureus did not enhanced the biofilm thickness of C. albicans and its PGE 2 synthesis in dual biofilms compared with mono-microbial C. albicans biofilms, although bacterial peptidoglycan-derived molecules have been shown to promote C. albicans hyphal growth [66,67]. Thus, in mixed biofilms the impact of S. aureus to C. albicans remained unclear.

Conclusion
Our findings indicate that PGE 2 is the key molecule stimulating the growth and biofilm formation of S. aureus in dual S. aureus/C. albicans biofilms, although subinhibitory farnesol concentrations may also support this effect. Candidal PGE 2 may exhibit a dual effect in S. aureus/C. albicans polymicrobial biofilms, first, by promoting fungal hyphal formation and second, by providing a proper substratum for the proliferation of S. aureus. Further characterization of the intricate interaction between these pathogens is warranted, as it may aid in the design of further therapeutic strategies against polymicrobial biolfilm infections.