Introduction

In healthy individuals, Candida albicans is a common commensal of the skin, the oropharynx, the gastrointestinal and the vaginal tract1. The fungus can however shift to an opportunistic pathogen and cause infections, ranging from superficial infections of the mucosa to invasive, life-threatening disease. The most prevalent C. albicans infection is vulvovaginal candidosis (VVC), which affects approximately 75% women at least once during their lives, with reported risk factors including diabetes and being part of a couple2,3. Other Candida infections, including oral and oropharyngeal candidosis, are common in immunocompromised individuals1,4,5. Standard treatment of C. albicans infections consists of azole therapy, however the development of azole resistance in Candida species is rising and worrisome6. Furthermore, these standard treatments can also cause side effects, including skin irritation, redness, gastrointestinal disorders and hepatotoxicity7. In addition, treated Candida infections are often followed by relapses; for instance, 103–172 million women annually suffer from recurrent VVC after an initial infection, with more than four confirmed VVC episodes per year8. This is possibly due to aggravated microbiome dysbiosis by azole use, and we recently observed a trend of reduced endogenous lactobacilli after fluconazole treatment9. Altogether, this signifies the need for novel therapeutic strategies.

An increasing number of clinical studies show the safety and efficacy of probiotic microorganisms to prevent or treat VVC10. Different study designs have been explored, with different follow-up durations, different routes of probiotic administration, or treatment combinations (the probiotic alone, a mixture of probiotics or in an adjuvant setting with azole treatment), mostly with a focus on oral intake. Several studies show that oral (as a food supplement) and vaginal (generally as a drug, with a more complex regulatory route) intake of different probiotic species and strains is associated with improved clinical signs11,12,13. In addition, no side effects have been observed or reported upon probiotic use in healthy women14 or women suffering from VVC11,15,16.

Despite the promising results from several trials with specific probiotic microorganisms, their multifactorial biological activity is often not fully understood. Yet, substantiation of the biological activity underlying probiotic efficacy can refine strain selection, improve combinations of different microbial strains based on complementary or synergistic properties, allow for patient stratification between responders and non-responders and facilitate regulatory approval. Several studies have evaluated possible mode of actions of individual Saccharomyces and Lactobacillaceae strains such as Lactobacillus gasseri JCM 1131, Lactobacillus crispatus JCM 1185, Lacticaseibacillus rhamnosus GR-1 and Limosilactobacillus reuteri RC-14 in vitro. These include the production of antimicrobial molecules and their direct growth-inhibitory and biofilm-inhibitory activity17,18, anti-adhesive capacity through co-aggregation or site exclusion19,20, and interference in hyphae formation, which is an important step in the infectious process and key for the ability of C. albicans to invade mucosal epithelial cells9,21,22. We have recently discovered that lactobacilli occur in modules or guilds of interacting bacteria in the vagina, with Lactobacillus crispatus, Lactobacillus jensenii and Limosilactobacillus taxa co-occurring23. However, a functional role for these modules, the interacting taxa, with Limosilactobacillus taxa in particular, is currently underexplored. Similarly, while probiotic strains are routinely combined during clinical evaluation against VVC24, their interacting effects (such as lack of antagonism and possibility for synergism) are often not explored during the in vitro evaluation of mode-of-actions.

In this study, we aimed to explore combinations of lactobacilli (in particular Lim. fermentum) and the yeast Saccharomyces cerevisiae CNCM I-3856 to target Candida infections in the vagina and other mucosal surfaces such as the oral cavity. S. cerevisiae CNCM I-3856 was previously reported to be effective against Candida infections in vitro and in animals25,26. S. cerevisiae CNCM I-3856 was also shown to be able to migrate from the gut to the vagina27, making a oral food supplementation feasible to target the vagina. However, this yeast does not always result in long-term colonization of the vagina in healthy subjects after oral administration27. In addition, S. cerevisiae is not a dominant member of the vaginal microbiota9,16. Therefore, here we investigated the multifactorial biological activity of probiotic S. cerevisiae CNCM I-3856 in combination with Lactobacillaceae strains from vaginal and food origin against C. albicans. We tested three vaginal Lactobacillaceae species, namely Lactobacillus crispatus, Lactobacillus johnsonii and Lacticaseibacillus rhamnosus, which persist naturally or upon administration in the vagina, at least temporarily9,16,23. In addition, we tested multiple Lim. fermentum strains isolated from food sources suitable for oral application. Sourdough is characterized by a microbial ecosystem comprised of lactic acid bacteria and yeasts that undergo beneficial metabolic interactions with each other28. Thus, sourdough lactobacilli isolates were included because of less risk of undergoing antagonistic interactions with S. cerevisiae CNCM I-3856. Another important reason to include lactobacilli isolates from sourdough and vegetable fermentations was that these genera and species of lactobacilli are naturally found in the human gastrointestinal tract29,30, while we have also recently demonstrated that several of these taxa can also be routinely isolated from the human vagina23. We evaluated potential synergistic and absence of antagonistic effects between the yeast and the selected bacterial species for the development of probiotic combinations of the well-studied S. cerevisiae CNCM I-3856 with at least one Lactobacillaceae strain, and to better understand the interactions of the probiotic S. cerevisiae CNCM I-3856 with lactobacilli potentially found in the gastrointestinal and vaginal microbiome.

Material and methods

Microbial strains used in this study

The clinical isolate C. albicans SC531431 was used as a pathogenic strain. The probiotic yeast S. cerevisiae CNCM I-385615,25,26,27 was obtained from a commercially available probiotic product (Gnosis by Lesaffre, Marcq-en-Baroeul, France) and used either as overnight culture, cell-free culture supernatant or commercial dried powder26. A total of 19 Lactobacillaceae strains were selected as explained above in the “Introduction” section, and used for the analyses as detailed in Table 1 either as overnight culture or cell-free culture supernatant. We used live lactobacilli and not a powder formulation, as this is how they are found in the gastrointestinal tract and the vagina. Yeasts were grown in yeast extract peptone dextrose (YPD) broth (Carl Roth), while lactobacilli were grown in De Man, Rogosa and Sharpe (MRS) broth (Difco). Growth of microorganisms was measured using spectrophotometer at 600 nm.

Table 1 Lactobacilli isolates used in this study and their origin.
Full size table

Spot assay for monitoring growth inhibition of C. albicans

Spot assay against C. albicans was performed as previously described9 with minor modifications. Briefly, 2 µL of overnight cultures of lactobacilli or S. cerevisiae CNCM I-3856 were spotted on solid agar at set distances from each other (1 cm, 1.5 cm or 2 cm, allowing to evaluate individual and possible synergistic effects). After 24 h of spots incubation, an overlay of a soft YPD agar (0.5% agar) (Carl Roth) containing C. albicans SC5314 at 4 × 107 CFU/mL was poured over the spots. After overnight incubation at 37 °C in aerobic conditions, growth of C. albicans SC5314 was evaluated, and inhibition zones were measured in mm.

Time course analysis of C. albicans growth inhibition and of S. cerevisiae and lactobacilli growth

Time course analysis of C. albicans growth inhibition was performed as previously described19 with minor modifications. Briefly, spent culture supernatants of overnight Lactobacillaceae and S. cerevisiae CNCM I-3856 cultures was obtained by centrifugation (10 min, 2000 g) and filtration with a 0.22 μm sterile syringe filter (VWR). The following supernatants from overnight cultures were tested against C. albicans SC5314: (1) supernatant of Lactobacillaceae as such in MRS broth; (2) supernatant of S. cerevisiae as such in YPD broth; (3) mixture of Lactobacillaceae and S. cerevisiae CNCM I-3856 supernatants in MRS and YPD, respectively. The supernatant was supplemented to fresh growth medium of C. albicans in a 1:10 ratio for a total volume of 200 µL and added to a flat-bottom 96-well plate (VWR). Each well was inoculated with an overnight culture of C. albicans SC5314 at 1% (v/v). Conditions with C. albicans as such, C. albicans with MRS at pH 4 and C. albicans with miconazole (80 µg/mL) were included as control conditions. Growth of C. albicans in continuous shaking conditions and 37 °C was evaluated over 24 h by measuring the optical density at 600 nm every 30 min in each well of the 96-well plate with the Synergy HTX multimode reader.

In addition, the same methodology was used to test the effects of vaginal Lactobacillaceae supernatants on the growth of S. cerevisiae CNCM I-3856, and the effects of the S. cerevisiae CNCM I-3856 supernatant on the growth of vaginal Lactobacillaceae strains.

C. albicans co-aggregation/agglutination and binding assay with S. cerevisiae and lactobacilli

Co-aggregation/agglutination with C. albicans yeast cells

The non-fluorescent co-aggregation assay was based on26 with several adjustments. Overnight C. albicans SC5314 cultures were washed twice (1500 × g, 10 min) and resuspended in phosphate buffered saline (PBS) to obtain a 1% (w/v) cell suspension. S. cerevisiae CNCM I-3856 from commercial dried powder was also resuspended in PBS to obtain a 1% w/v cell suspension. Lactobacillaceae cultured overnight were adjusted to a concentration of 1010 CFU/mL. After vortexing, 50 µL of C. albicans suspension, 50 µL of S. cerevisiae suspension and 25 µL of lactobacilli were added to the wells of a U-bottom 96-well plate (VWR). All Lactobacillaceae and S. cerevisiae were also tested as such with C. albicans. The plate was incubated at room temperature with gentle shaking. After 10 min and 1 h, the co-aggregation rate was microscopically evaluated using the Olympus CX41 light microscope and Olympus U-CMAD3 camera. Scores from 0 to 4 were given for each condition, as described by26, with a score of 0: no aggregation; 1: aggregates with small clusters; 2: aggregates with larger numbers of yeasts; 3: clumps visible with the naked eye containing large numbers of yeast cells; 4: maximum score for large clumps visible with the naked eye in the well center. For conditions where even higher aggregation was observed, we rationalized that a higher score was needed and indicated this with 4+.

Co-aggregation/agglutination with C. albicans after hyphae induction

Furthermore, fluorescent labelling of C. albicans SC5314 and S. cerevisiae CNCM I-3856 partially based on Pericolini et al.26 was performed to assess their binding/co-aggregation. Overnight culture pellet containing 1 × 107 CFU of C. albicans was resuspended in 250 µL CalcoFluor White (CFW) (VWR) and stained with 0.5 mg CFW/mL. S. cerevisiae CNCM I-3856 at containing 1 × 107 CFU from commercial dried powder was resuspended in 250 µL Fluorescein isothiocyanate (FITC) solution and stained with FITC (F7250-1G, Sigma) at a concentration of 0.4 mg/mL. Labelling was performed at room temperature for 1 h at 250 rpm. After centrifugation for 10 min at 4000 g, the pellet was washed using PBS and resuspended in 1 mL to obtain a concentration of 4 × 107–108 CFU/mL for C. albicans and S. cerevisiae CNCM I-3856. To assess potential binding based on fluorescence in mixtures of C. albicans with S. cerevisiae CNCM I-3856, mixtures were made by adding 12.5 µL of the C. albicans suspension to 12.5 µL of fetal calf serum (FCS) (Gibco) and 100 µL YPD broth in a 96 U-bottom well plate. After 3 h of incubation under non-shaking conditions at 37 °C 50 µL of the FITC-stained S. cerevisiae CNCM I-3856 suspension was added to the wells with C. albicans and incubated at room temperature with gentle shaking (200 rpm). After 10 min or 1 h of incubation, 2 µL of the mixtures was used for microscopic evaluation by checking binding between cells or cells and hyphae, and cluster formation between C. albicans SC5314 (blue) and S. cerevisiae CNCM I-3856 (green). Fluorescence microscopy images were recorded with the Leica DMi8 fluorescence microscope.

Inhibition of C. albicans hyphae formation by S. cerevisiae and lactobacilli

Two protocols were implemented to assess inhibition of C. albicans SC5314 hyphae formation by S. cerevisiae CNCM I-3856 cell-free culture supernatant and Lactobacillaceae or their mixtures based on previously developed protocols by Allonsius et al.19,21. In the first protocol not implementing fluorescent labeling, mixtures of 106 CFU/mL of C. albicans SC5314 with cell-free culture supernatant of S. cerevisiae CNCM I-3856, with Lactobacillaceae or both were made by adding 50 µL of suspension of each tested microorganism to 125 µL of FCS, supplemented with YPD broth to a total volume of 500 µL. For each condition three technical repeats were included. After 3 h of incubation in non-shaking conditions at 37 °C, 2 µL of the mixtures was used for microscopic evaluation by counting the number of yeast cells and the number of cells forming hyphae. At least 100 cells were counted and ratios of hyphae to yeast cells were calculated and normalized to C. albicans as such with FCS. For normalization, the average ratio of hyphae to yeast cells counted for C. albicans as such (after incubation with FCS, “C. albicans control”) was first calculated. Afterwards, the number of yeast cells and number of hyphae were counted for each repetition of each conditions and the ratios were calculated. These ratios were then divided by the average ratio of hyphae to yeast cells counted for C. albicans as such (after incubation with FCS, “C. albicans control”).

The second protocol implemented fluorescent labeling to test the effects of live S. cerevisiae CNCM I-3856 from commercial dried powder and Lactobacillaceae or their mixtures for inhibition of C. albicans SC5314 hyphae formation. The two best performing strains from the first hyphal inhibition experiment, namely L. fermentum LS4 and LS5, were selected for this experiment to be used as such or in combination with live S. cerevisiae CNCM I-3856. S. cerevisiae CNCM I-3856 powder was reconstituted in YPD broth after weighing the dried powder 0.1% w/v, and the resuspended powder was shaken for 15 to 20 min at 100 rpm. The concentrations of overnight cultures of Lactobacillaceae and yeast were subsequently adjusted to 109 CFU/mL and 107 CFU/mL, respectively. Fluorescent staining of C. albicans SC5314 and S. cerevisiae CNCM I-3856 was performed as described above for the co-aggregation assay. Inhibition of C. albicans hyphae formation by S. cerevisiae CNCM I-3856 as commercial powder, overnight culture or supernatant was tested. To assess hyphae formation in mixtures of C. albicans with S. cerevisiae and/or lactobacilli L. fermentum LS4 or L. fermentum LS5, mixtures were made by adding 50 µL of suspension of each tested microorganism to 125 µL of FCS, supplemented with YPD broth until 500 µL. For each condition three or four technical repeats were included in a 24 well plate. After 3 h of incubation at 37 °C to allow C. albicans hyphae formation, 2 µL of the mixtures was used for microscopic evaluation by counting the number of yeast cells and the number of cells forming hyphae. At least 100 cells were counted and ratios of hyphae to yeast cells were calculated and normalized to C. albicans as such with FCS. Fluorescence microscopy images were recorded with the Leica DMi8 fluorescence microscope.

Statistical analysis

Significant differences between all tested conditions were evaluated in GraphPad Prism version 9.2.0 for Windows (GraphPad Software, San Diego, California USA, www.graphpad.com). Statistical testing was performed using one-way ANOVA (when testing one experimental factor, for example effects of lactobacilli strains) or two-way ANOVA (when testing several experimental factors, for example effects of lactobacilli as such or with S. cerevisiae) with Dunnett's multiple comparisons test to identify pairs of conditions with significant differences compared to the control group; p-values < 0.05 were considered significant. Experimental conditions were tested at least in triplicates.

Results

Specific strains of lactobacilli as such or in combination with S. cerevisiae CNCM I-3856 inhibit growth of Candida albicans

A total of 19 lactobacilli isolates of food (sourdough, and fermented plants and vegetables) and vaginal origin were selected for testing against C. albicans SC5314 (Table 1).

First, we assessed whether live metabolically active lactobacilli or S. cerevisiae CNCM I-3856 could inhibit the growth of C. albicans by performing spot assays. Almost half of the lactobacilli tested (9/19; 47%) showed growth inhibitory activity against C. albicans, with inhibition zones ranging between 1 and 3 mm (Fig. 1A). The largest inhibition zone of 3 mm was observed for L. rhamnosus AMBV083. Also L. plantarum LS2, L. fermentum LS5, L. paracasei LS14 and L. carotarum AMBF275 showed a strong inhibitory capacity against C. albicans with approximately 2 mm inhibition zones. However, no specific growth inhibition of C. albicans by S. cerevisiae CNCM I-3856 was detected.

Figure 1
figure 1

(A) Growth inhibition zones in the spot assay against C. albicans SC5314 measuring growth inhibition of C. albicans by cultures (spots) of live S. cerevisiae CNCM I-3856 and the tested lactobacilli as such. (B) Growth inhibition of C. albicans by supernatant of S. cerevisiae, the lactobacilli as such, or the mixture of supernatant of the lactobacilli and supernatant of S. cerevisiae. MRS medium at pH 4 was used as an acidic control condition, while miconazole was used as antifungal control. Bars represent the calculated area under the C. albicans growth curves and data are expressed as mean ± standard deviation (SD). Significant differences compared to C. albicans control (dotted line) are shown (*p < 0.05, **p < 0.001, ***p < 0.001, ****p < 0.0001).

Full size image

Subsequently, we evaluated whether secreted metabolites of lactobacilli, secreted metabolites of S. cerevisiae or their combination inhibited the growth of C. albicans in a time-course assay. Cell-free culture supernatants of 11 out of 19 tested lactobacilli inhibited the growth of C. albicans by at least 5% (Fig. 1B). This was not the case for MRS at pH 4, which was used as a control, suggesting an additional effect of lactobacilli metabolites in addition to the acidic environment. The combinations of lactobacilli with the supernatant of S. cerevisiae CNCM I-3856 led to similar C. albicans growth inhibition, without significant differences compared to the supernatants of lactobacilli as such. Supernatant of S. cerevisiae alone did not inhibit C. albicans growth under the tested conditions (Fig. 1B). Importantly, the supernatant of S. cerevisiae CNCM I-3856 did not lead to significant longitudinal inhibition of the growth of the vaginal lactobacilli isolates (Fig. S1B). Similarly, the culture supernatants of the vaginal lactobacilli did not significantly reduce the growth of S. cerevisiae CNCM I-3856 (Fig. S1A).

Altogether, L. plantarum LS1, L. plantarum LS2, L. plantarum LS3, L. plantarum LS7 and L. rhamnosus AMBV083 showed the strongest antimicrobial effects on C. albicans in both assays performed here. The longitudinal effects of these best performing lactobacilli strains’ supernatants on the growth curves of C. albicans are shown in the Supplementary Material (Fig. S2).

S. cerevisiae CNCM I-3856 as such and in combination with lactobacilli promotes Candida albicans aggregation

Another possible biological mechanism of S. cerevisiae and lactobacilli to prevent Candida adhesion and subsequent infection is by aggregating with C. albicans, because aggregation promotes closer contact with potential antimicrobial metabolites and can block adhesion to target host sites by C. albicans. Here, we assessed the cluster formation of C. albicans with lactobacilli, with S. cerevisiae or in combination after short-term co-incubation (10 min) by microscopic evaluation.

S. cerevisiae aggregated with the yeast form of C. albicans after 10 min of incubation (Fig. 2A; Supplementary Table S1). The majority (13/19; 68%) of the lactobacilli also aggregated with the yeast form of C. albicans cells after 10 min of incubation with a score of 1 out of 4: L. plantarum LS3, L. fermentum LS4, L. paracasei LS12, L. paracasei LS13, L. paracasei LS14, L. crispatus AMBV012, L. johnsonii AMBV023, L. rhamnosus AMBV083 and L. reuteri AMBF471. L. carotarum AMBF275 showed the strongest aggregation with C. albicans with a score of 2 out of 4.

Figure 2
figure 2

(A) Aggregation scores of the yeast form of C. albicans SC5314 cells with lactobacilli as such, or in co-culture with dried S. cerevisiae CNCM I-3856 after 10 min of incubation; (B, C) aggregation of S. cerevisiae CNCM I-3856 (in green, stained with FITC) with C. albicans SC5314 after hyphae induction (in blue, stained with CFW) after 10 min (B) or 1 h (C) of co-incubation. The clusters formed were scored from 0 (no clusters) to 4+ based on their size according to26.

Full size image

Remarkably, the combination of S. cerevisiae and lactobacilli gave a strong aggregating effect after 10 min of incubation, which was observed for all lactobacilli. While the aggregating effect was clear from microscopic evaluation also after a longer incubation of 1 h (Supplementary Fig. S3), the synergistic effect of S. cerevisiae and lactobacilli on the yeast form of C. albicans aggregation was not strengthened by the longer incubation (Supplementary Table S2).

Additionally, a visual indicative assessment of S. cerevisiae CNCM I-3856 binding to Candida albicans after hyphae induction was performed using fluorescently stained S. cerevisiae CNCM I-3856 and Candida albicans. Aggregation of the hyphal form of C. albicans with S. cerevisiae CNCM I-3856 was visually confirmed to occur via fluorescent microscopy with differentially stained cells (Fig. 2B) after both 10 min and 1 h of co-incubation.

Lactobacilli as such or in combination with S. cerevisiae CNCM I-3856 inhibit hyphae formation by Candida albicans

Lastly, we tested the effects of S. cerevisiae and lactobacilli on hyphae formation of C. albicans necessary for this pathogen to invade mucosal epithelial cells9,19,22. In the first set of experiments, the effects of live lactobacilli as such and in combination with cell-free culture supernatant of S. cerevisiae were tested for their ability to inhibit hyphae formation by C. albicans (Fig. 3 and Supplementary Fig. S4). Significant inhibition of hyphae formation was demonstrated for L. fermentum LS4 (by 62 ± 14% compared to 100% Candida control), L. fermentum LS5 (by 78 ± 14%), L. paracasei LS6 (by 33 ± 20%), L. johnsonii AMBV083 (by 32 ± 14%) and L. reuteri AMBVF471 (by 61 ± 24%) as such. This was also the case for combinations with supernatant of S. cerevisiae CNCM I-3856, especially for L. fermentum LS4 (inhibition of hyphae formation by 78 ± 17% compared to 100% Candida control).

Figure 3
figure 3

C. albicans SC5314 hyphal formation inhibition during co-incubation with lactobacilli as such or combined with cell-free culture supernatant of S. cerevisiae CNCM I-3856. A concentration of 1/4 FCS was used to induce hyphae formation. Results are represented as ratio against a total number of 100 yeast cells and normalized against C. albicans as such. Data are expressed as mean ± SDs. Significant differences compared to C. albicans control are shown with an asterisk (* p < 0.05, ** p < 0.001, **** p < 0.0001).

Full size image

Based on the results of Fig. 3, the two best performing strains from the first hyphal inhibition experiment showing the strongest inhibition of C. albicans hyphae formation, namely L. fermentum LS4 and LS5, were selected to be further evaluated in a second set of experiments. In these experiments, they were combined with fluorescently labelled live S. cerevisiae CNCM I-3856 from dried powder and tested for their ability to inhibit hyphae formation by C. albicans. Significant inhibition of C. albicans hyphae formation was observed both by L. fermentum LS4 (by 92.5 ± 4% compared to 100% Candida control) and L. fermentum LS5 (by 88.5 ± 4%) as such and their combination with live S. cerevisiae CNCM I-3856 (Fig. 4).

Figure 4
figure 4

(A) Hyphal induction of C. albicans SC5314 during co-incubation with live S. cerevisiae CNCM I-3856 together with L. fermentum LS4 or LS5, or with L. fermentum LS4 or LS5 as such. A concentration of 1/4 FCS was used to induce hyphae formation in C. albicans (CA + FCS). S. cerevisiae CNCM I-3856 powder was used together with C. albicans with FCS (CA + FCS + SC), or also with the different lactobacilli (CA + FCS + SC + LS4; CA + FCS + SC + LS5). Also conditions with C. albicans and lactobacilli without S. cerevisiae were included (CA + FCS + LS4; CA + FCS + LS5). Results are represented as ratio against a total number of 100 yeast cells and normalized against C. albicans as such. Data are expressed as mean ± SD. Significant differences compared to C. albicans with FCS conditions (CA + FCS) are shown with an asterisk (* p < 0.05, **** p < 0.0001); (B) fluorescence microscopy examples from the experiment on hyphal induction of C. albicans during co-incubation with live S. cerevisiae CNCM I-3856 together with L. fermentum LS4 or LS5. A concentration of 1/4 FCS was used to induce hyphae formation in C. albicans (CA + FCS), while C. albicans (CA) as such had no FSC added. S. cerevisiae CNCM I-3856 powder was used together with C. albicans with FCS (CA + FCS + SC), or also with the different lactobacilli (CA + FCS + SC + LS4; CA + FCS + SC + LS5; CA + FCS + SC + LS7).

Full size image

Discussion

In this study, we investigated the in vitro efficacy of probiotic Saccharomyces cerevisiae CNCM I-3856 and Lactobacillaceae strains against C. albicans SC5314 through different potential types of biological activity. We showed that lactobacilli alone or in combination with S. cerevisiae CNCM I-3856 lead to growth inhibition, agglutination and hyphae inhibition of C. albicans in vitro, thus acting against at least three key aspects involved in C. albicans pathogenesis in vivo (Fig. 5). The significance of the observed lactobacilli-mediated effects was species-specific (e.g., for hyphae inhibition) or even strain-specific (e.g., for C. albicans growth inhibition) for each of the discussed types of biological activity summarized in Fig. 5. The tested L. fermentum strains most efficiently inhibited hyphae formation, but also other strains showed significant anti-hyphae activity, including L. paracasei and L. reuteri strains. The tested L. carotarum strain scored the highest capacity for C. albicans aggregation. Candida growth inhibition in the tested conditions was modest, with predominantly L. plantarum and L. rhamnosus demonstrating the highest inhibition of C. albicans growth via secreted metabolites. For the species L. paracasei, we found that the activity was strongly strain-specific, with strains showing no growth inhibition (LS11, LS12) and strains that did (LS13, LS14).

Figure 5
figure 5

Postulated biological activity through which S. cerevisiae CNCM I-3856 and Lactobacillaceae strains inhibit C. albicans. The contribution of each mechanism can vary depending on the microbial strains used and the in vivo context (e.g., the microbiome). Created with BioRender.com.

Full size image

Synergistic effects between lactobacilli and the yeast S. cerevisiae CNCM I-3856 against vaginal C. albicans in our work were observed at the level of agglutination in the tested in vitro set-ups. A tendency towards a stronger anti-hyphae effect of L. fermentum LS4 combined with S. cerevisiae CNCM I-3856 compared to L. fermentum LS4 was also observed in some experiments, however the effect L. fermentum LS4 and LS5 as such against hyphae was so pronounced that the demonstration of synergy with S. cerevisiae CNCM I-3856 was complicated in our in vitro experimental set-ups. Of note, other ecological mechanisms active in vivo and synergistic effects with other microbiome members that can further boost the efficacy of the S. cerevisiae CNCM I-3856 and lactobacilli are not excluded. Such synergistic in vivo effects could explain the more promising results against Candida hyphae formation observed in a mouse model of VVC upon S. cerevisiae CNCM I-3856 administration, in addition to differences in experimental set-ups26. Likewise, a previous in vivo clinical study in women demonstrated that orally administered S. cerevisiae CNCM I-3856 can serve as an effective adjuvant therapy to conventional VVC treatment through inhibiting vaginal Candida proliferation15.

Importantly, we did not observe significant antagonistic activity between S. cerevisiae CNCM I-3856 and the tested lactobacilli regarding C. albicans inhibition, highlighting that they can be applied together without loss of activity. Furthermore, it is plausible that S. cerevisiae CNCM I-3856 will not have major antagonistic effects with the lactobacilli comprising the natural vaginal microbiome, as we have shown for the vaginal lactobacilli in this paper, and recently demonstrated in healthy women14. This is also relevant for oral S. cerevisiae CNCM I-3856 administration, as it has previously been demonstrated that after its oral administration in women S. cerevisiae CNCM I-3856 can be recovered from vaginal samples of at least 18% to 21% of participants, depending on the analysis14,27. Also in the gut, a combined effect of S. cerevisiae CNCM I-3856 and strains of L. plantarum, L. rhamnosus and L. fermentum strains can be hypothesized, as these lactobacilli are naturally present in fermented foods29 and, consequently, in the human gut depending on the individual diet30.

Our results provide insight in the biological activity underlying the effects observed in vivo in the studies of Pericolini et al.26 and Cayzeele-Decherf et al.15. The action of lactobacilli and S. cerevisiae CNCM I-3856 in vivo is likely multifactorial, as summarized in Fig. 5. While in this study we have not explored which exact molecules of lactobacilli are responsible for Candida hyphae or growth inhibition, we have previously already demonstrated that one of the most prominent surface molecules of lactobacilli, exopolysaccharides (EPS), could inhibit growth and hyphae formation19, but the activity was even higher for the peptidoglycan hydrolase major secreted protein 1 (Msp1) with chitinase functionality produced by Lacticaseibacillus strains21. Related to direct antimicrobial interactions, the main secreted molecules in lactobacilli are d- and l-lactic acid, which have been linked to inhibition of C. albicans virulence by multiple studies19,21,36. In addition to their impact on the different virulence factors of C. albicans (growth, adhesion via aggregation and hyphae formation), the lactobacilli and S. cerevisiae CNCM I-3856 cells may also interact with each other and with C. albicans via other mechanisms, which might become more prominent in vivo. For example, ecological interactions such as nutrient competition and metabolite production were not explored in this study, yet competition for nutrients has been described between S. cerevisiae and Candida37 and thus has been included in potential mechanisms in Fig. 5. Finally, another potential mechanism might be through microbiome modulation. Although our recent exploratory, randomized, double-blind, placebo-controlled clinical study (n = 60) only showed limited effect of probiotic supplementation of S. cerevisiae CNCM I-3856 on the fungal and bacterial community of healthy women14, the in vitro results of the current manuscript suggest that the effects of S. cerevisiae CNCM I-3856 in combination with specific lactobacilli providing optimal probiotic action are promising to further investigate in women suffering from vaginal infections.

Several limitations of our study lie in the fact that specific in vitro assays were implemented, which do not always reflect in vivo conditions. First, the standard deviation per condition was rather high when in vitro results of several experimental repeats were combined, which can be explained by the intrinsic variability of tripartite biological assays combining three microorganisms or their supernatants (Candida, Saccharomyces and lactobacilli). Indeed, some experimental groups presented values greater than 100% in the hyphae formation experiment relative to the average of the Candida control condition. Another example is that the supernatants of lactobacilli and S. cerevisiae used in the anti-Candida growth assays had to be diluted, possibly reducing their biological activity. In our study, careful statistical testing was implemented to elucidate whether the observed effects could have occurred by chance. Another limitation of in vitro assays is that dedicated choices have to be made regarding the biological mode of action to focus on, and in future work also the effects of S. cerevisiae CNCM I-3856 and lactobacilli against C. albicans biofilms should be explored. Second, the probiotic activity is caused not only by individual probiotic species as they are studied in vitro, but also by their multi-microbial interaction with resident microbial communities in vivo, making it important to take whole microbial communities into consideration38. To address the study limitations, in the future we suggest to implement the tested microorganisms in innovative assays allowing to better mimic the human body environment (e.g., organoid systems), monitor relevant health read-outs in vitro (e.g. vagina-on-a-chip) and potentially add synthetic microbial communities or whole microbiomes resident in the human body. Furthermore, additional in vivo studies should be conducted with detailed microbiome and metabolomics read-outs to better understand the role of S. cerevisiae CNCM I-3856 and lactobacilli against C. albicans as part of an integrated system in the female gut and vagina.

Taken together, our results pave the way for different probiotic combinations or consortia, for example opening up possibilities for combining specific L. plantarum, L. rhamnosus, L. fermentum and L. carotarum probiotics with S. cerevisiae CNCM I-3856. On the other hand, our findings on the strain-specific effects of lactobacilli in combination with S. cerevisiae CNCM I-3856 are especially important considering the knowledge that the vaginal microbiome of women is dominated by different species of lactobacilli. Our recent results from the Isala study in a large Belgian cohort show that L. plantarum, L. rhamnosus and L. fermentum are naturally present in the female vagina at high prevalence, but not high relative abundance23, suggesting that S. cerevisiae CNCM I-3856 might be more effective in some women due to their natural gut or vaginal microbiome composition. In line with the Isala study, our results further underline the potential keystone function of Limosilactobacillus to promote health in the vaginal niche23. Considering that the vaginas of healthy women are colonized with different lactobacilli and it is not yet clear how differences in dominant lactobacilli can be explained, personalized therapies with different combination of probiotic lactobacilli and S. cerevisiae CNCM I-3856 might be needed for positive clinical outcomes.

Conclusion

We have demonstrated the potential multifactorial biological activity of specific Lactobacillaceae strains alone or combined with the probiotic S. cerevisiae CNCM I-3856 against C. albicans in vitro, which was more pronounced for anti-hyphae and agglutination action and was less significant for growth inhibition under the tested conditions. Our results show strain-specific anti-Candida modes of action of lactobacilli and no significant antagonistic effects between lactobacilli and S. cerevisiae CNCM I-3856, which can help inform the development of effective combinations of lactobacilli and yeast probiotics. Furthermore, the observed differences between strains of lactobacilli highlight the potential importance of the endogenous female microbiome in modulating the action of S. cerevisiae CNCM I-3856 against C. albicans, which can help in the development of efficient personalized choices of probiotic therapies, although this further needs to be studied in vivo.