Saccharomyces cerevisiae show low levels of traversal across human endothelial barrier in vitro

Background: Saccharomyces cerevisiae is generally considered safe, and is involved in the production of many types of foods and dietary supplements. However, some isolates, which are genetically related to strains used in brewing and baking, have shown virulent traits, being able to produce infections in humans, mainly in immunodeficient patients. This can lead to systemic infections in humans. Methods: In this work, we studied S. cerevisiae isolates in an in vitro human endothelial barrier model, comparing their behaviour with that of several strains of the related pathogens Candida glabrata and Candida albicans. Results: The results showed that this food related yeast is able to cross the endothelial barrier in vitro. However, in contrast to C. glabrata and C. albicans, S. cerevisiae showed very low levels of traversal. Conclusions: We conclude that using an in vitro human endothelial barrier model with S. cerevisiae can be useful to evaluate the safety of S. cerevisiae strains isolated from foods.


Introduction
Saccharomyces cerevisiae is generally considered safe, and is involved in the production of a variety of foods and dietary supplements. Several types of food and beverage still contain viable yeast cells [1][2][3][4][5] . However, in the last years human infections with Saccharomyces cerevisiae have increased 6-8 . Consequently, S. cerevisiae is considered an emerging pathogen 9-11 . Different parts of the body can be affected in immunocompromised [12][13][14][15] and healthy patients [16][17][18] . The potential virulence of this yeast has been analysed with different methods in vitro [19][20][21][22] and in vivo 23-27 , for example by measuring epithelial barrier traversal 28 . These reports have suggested that certain strains can cause disease and death in murine models. However, the bio-therapeutic agent Ultralevure (S. cerevisiae var. boulardii) and other supplements are consumed in high doses, ranging from 10 7 to 10 10 live yeast cells per day and for long periods.
The study of yeast virulence includes studying their behaviour when they encounter endothelial barriers. Opportunistic pathogenic yeasts such as C. glabrata and C. albicans are able to pass the intestinal barrier 29,30 and generate systemic infections 31-33 . Also, C. albicans can cross the blood-brain barrier (BBB) to reach the brain 34,35 . Regarding S. cerevisiae, infections after oral ingestion 16 or digestive translocation 12,14,36 show that it can reach brain in murine models 25 . However, few studies have investigated the behaviour of S. cerevisiae when they reach endothelial barriers 28 .

Methods
Yeast strains and growth media The yeast strains are described in Table 1. Strains were propagated in YPD media (1% glucose, 1% BactoPeptone, 0.5% yeast extract) for 24 h at 30°C.
Trans-epithelial electrical resistance (TEER) assay HUVEC cells (1×10 5 cells/cm 2 ) were seeded on Transwell® filter inserts (8 μm, Corning Incorporated) in 24-well plates (Corning Incorporated). A volume of 200 μL cell growth medium was added to the apical compartment and 1250 μL to the basolateral compartment. The TEER was measured using the Millicell-ERS Electrical Resistance System (Millipore). The net value of the TEER (Ωcm 2 ) was corrected for background resistance by subtracting the contribution of the cell-free filter and the medium (110 Ωcm 2 ). The TEER was measured before the addition of yeasts.
Determination of permeability coefficient 1 μg/mL of fluorescein (Sigma) was added to the media in the apical compartment of the transwell, with or without

Amendments from Version 1
In this new version, we have considered HUVEC cells as a model of endothelial cells instead of a blood brain barrier.

REVISED
established HUVEC monolayers, and fluorescence was measured over time in the media of the apical and basolateral compartment. The apparent permeability, Papp, was defined as (Hilgers et al., 1990): (ΔA R /Δt) is the rate of drug appearance in the receiver side, S is the surface area of the Transwell (0.33 cm 2 for Transwell® inserts (8 μm pore size, Corning) of 6.5-mm insert diameter), and C D,0 is the initial drug concentration in the donor side at time = 0. Values are expressed in cm/s.
Ability to cross the endothelial barrier HUVEC cells were seeded on Transwell® filter as described above. Yeasts grown overnight at 30°C in YPD were resuspended (10 6 cells mL -1 ) in the apical compartment and incubated at 37°C in a humidified atmosphere of 5% CO 2 and 95% air.
After 12 h, the basolateral compartment medium was replaced.
Colony forming units were counted in YPD plate triplicates after two days. Control wells used to evaluate yeast growth showed no significant growth after 12 h. Negative control HUVEC Transwells without adding cells were performed to control TEER stability across the experiment.

Evaluation of the endothelial barrier integrity
To establish an in vitro human endothelial barrier, we used HUVEC monolayers, a methodology that has been widely used 37,38 . Monolayers were formed in transwells and two different methods were used to determine the robustness, consistency and integrity of the barrier. First, we studied the TEER, indicative of physical separation. After seeding the HUVECs, TEER was measured and we observed increased values over time that were overcoming 450 Ωcm 2 , which correlates with the establishment of a monolayer barrier. Second, we studied the monolayer permeability. The value obtained was 1.82±0.13 (10 -6 cm/s) on average, which indicates an integral barrier with low permeability 39 .

Study of the ability of yeast species to cross the human endothelial barrier in vitro
To determine whether S. cerevisiae is able to cross the human endothelial barrier, we used an in vitro model of the endothelium with HUVECs 40 . The number of cells in the basolateral compartment was measured 12 hours after addition of S. cerevisiae, C. albicans and C. glabrata strains to the apical compartment ( Figure 1). The results showed that all yeast strains were able to cross the endothelial barrier. While elevated number of cells from C. glabrata and C. albicans strains were able to cross the endothelial barrier, S. cerevisiae values were low. Furthermore, while the S. cerevisiae control strain W303 showed the lowest levels of yeast transcytosis, the other opportunistic pathogenic strains presented higher levels.  To compare the different species, the average level of cell transcytosis for all strains of each species was calculated ( Figure 2 25 . When comparing to other well-known yeast pathogens such as C. glabrata and C. albicans, none of the S. cerevisiae strains were able to cross the endothelial barrier at high levels. Despite S. cerevisiae pathogenicity levels being lower than other opportunistic yeasts, we recommend the potential risk of new S. cerevisiae strains to be evaluated before using them in food production.

Competing interests
No competing interests were disclosed.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 1.