Forty-five C. albicans strains, which were clinically isolated from the vagina and provided by Microskylab Inc. (Tokyo, Japan), were used in this study. All 27 Lactobacillus strains were previously obtained from vaginal swabs of healthy Japanese women at Aichi Medical University (Matsumoto et al., 2018). These strains belonged to five species: L. crispatus, L. jensenii, L. gasseri, Lacticaseibacillus rhamnosus, and Limosilactobacillus vaginalis. The characteristics of these bacterial strains are listed in Supplementary Table 1.

A 96-well microtiter plate-based method was used in this study (Lohse et al., 2018). C. albicans strains were cultured for 24 h in Yeast Peptone Dextrose (YPD) agar (Difco, Detroit, MI, USA) at 30°C under aerobic conditions. A single colony was inoculated into the YPD broth medium and incubated overnight for 16 h at 30°C, accompanied with shaking at 160 rpm under aerobic conditions. Under these conditions, C. albicans strains grew to the budding yeast forms (blastospores). The cells were centrifuged at 3,500 ×g for 10 min and re-suspended in RPMI 1640 medium buffered with morpholinepropanesulfonic acid (MOPS) at a concentration of 10⁷ cells/mL, and 100 μL of the inoculum was seeded into a 96-well microtiter plate. The biofilms formed on the surface of the wells were gently washed twice with phosphate-buffered saline (PBS) after 48 h of incubation at 37°C. The yeast cells were not washed immediately after the initial adhesion, and thus, the final time point (48 h) reflects the total biomass that could not be initially adhered to. Crystal violet (CV) (Merck KGaA, Darmstadt, Germany) and water-soluble tetrazolium salts (WST-1) (TaKaRa, Shiga, Japan) were used in this study (Mukherjee et al., 2005; Weber et al., 2008). CV stained the whole biomass, including dead cells and polysaccharides, whereas WST was converted to a colored formazan in the presence of metabolic activity. To quantify the total biomass, washed biofilms were stained with 0.1% (w/v) CV solution for 1 min. Each well was washed twice with PBS and dried for 30 min. The bounded CV was eluted using 99.5% (v/v) ethanol. The burden of viable cells was estimated using WST-1 based on the reduction of tetrazolium salt. To each well, we added 100 μL of PBS and 10 μL of premixed WST-1, and the mixture was incubated at 37°C for 3 h under shade conditions. The absorbance (Abs) of CV and WST-1 was measured at 570 nm and 440 nm, respectively. Eight replicate wells were used for each strain, and experiments was repeated three times, independently.

ISOGEN II (Nippon Gene, Co., Ltd., Tokyo, Japan) was used for total RNA extraction from the C. albicans HB-10 strain. The RNA concentrations were measured using a Qubit^® RNA Assay Kit (Promega, WI, USA). To prepare complementary DNA, the PrimeScript™ RT reagent kit (TaKaRa, Shiga, Japan) was used in accordance with the manufacturer’s instructions. Moreover, qRT-PCR analysis was performed using TB Green^® Premix Ex Taq™ II (Tli RNaseH Plus) (TaKaRa, Shiga, Japan) in accordance with the manufacturer’s protocol. Briefly, PCR was performed in a reaction mixture of TB Green Premix Ex Taq II (2 ×) 12.5 µL, PCR forward primer 1 µL, PCR reverse primer 1 µL, and RNase free dH₂O 8.5 µL added to 2 µL of each reverse transcription reaction solution. Primers used in this study are listed in Supplementary Table 2. The amplification conditions were as follows: 40 cycles under heat treatment at 95°C for 30 s, heat denaturation at 95°C for 5 s, and annealing at 55°C for 30 s, which is the optimum temperature for the primer. Melting curves were used to verify the quality of qRT-PCR, and the fold expression was calculated using the delta-delta Ct method.

Cell-free culture supernatants were extracted from Lactobacillus species. A single strain each was inoculated in de Man, Rogosa, and Sharpe (MRS) broth (Merck KGaA, Darmstadt, Germany) and incubated at 37°C for 72-h under anaerobic conditions (10% H₂, 10% CO₂, and 80% N₂) in an anaerobic chamber. Growth at the sampling point (72-h) was determined by measuring the optical density (OD) at 600 nm using a microplate reader (SH-9000Lab, HITACHI). The culture medium was then centrifuged at 3,500 ×g for 10 min and filtered through a 0.22-μm membrane filter (Sarutorius AG, Gettingen, Germany). Each collected culture supernatant was stored at −80°C until use.

HPLC (SHIMADZU, Kyoto, Japan) equipped with a conductivity detector was used to measure levels of lactate and short-chain fatty acids, such as acetate, propionate, and butyrate, in culture supernatants, as previously described (Hagihara et al., 2021). Briefly, the mobile phase required 5 mM p-toluenesulfonic acid (KANTO Chemical, Tokyo, Japan). The reaction buffer was made of 5 mM p-toluenesulfonic acid, 100 μM ethylenediaminetetraacetic acid (KANTO Chemical, Tokyo, Japan), and 20 mM bis (2-hydroxyethyl) aminotris (hydroxymethyl) methane (Tokyo Chemical Industry, Tokyo, Japan). The flow rate, oven temperature, and detector cell temperature were set at 0.8 mL/min, 40°C, and 48°C, respectively. The samples contained in 1.0 mL disposable vials (SHIMADZU Co., Kyoto, Japan) were held at 4°C in a sample cooler (SHIMADZU, Kyoto, Japan), and 10 μL was applied to tandemly arranged two columns (SHIMADZU, Kyoto, Japan) to measure lactate levels. The calibration curve solution adjusted with lithium DL-lactate (FUJIFILM Wako Pure Chemical, Co., Ltd., Osaka, Japan) was dissolved in deionized water. The quantification analyses for HPLC were performed using LabSolutions version 5.90 (SHIMADZU Co., Kyoto, Japan), and the peak area was used as the signal intensity.

Hydrogen peroxide production was estimated using a hydrogen peroxide assay kit (ab102500, Abcam, MA, USA) according to the manufacturer’s instructions. The stored culture supernatant was neutralized to pH 7.0, and 100 μL of the adjusted supernatants were reacted for 10 min in the presence of horseradish peroxidase. Duplicate wells were measured for each sample with absorbance at 595 nm, and experiments were repeated three times, independently.

The pH of cell-free culture supernatants and buffered RPMI 1640 medium, supplemented with culture supernatant, were measured promptly using a glass electrode-style hydrogen-ion concentration meter (Laqua, Horiba, Ltd., Japan). MRS (8%) was added to RPMI instead of Lactobacillus supernatant to achieve final concentrations of lactate and hydrogen peroxide standard of 4–64 mM and 4 nM–64 nM, respectively.

C. albicans sHB-10 were cultured for 24 h in YPD agar at 30°C under aerobic conditions. A single colony was inoculated into the YPD broth medium and incubated overnight for 16 h at 30°C, accompanied with shaking at 160 rpm under aerobic conditions. The cells were centrifuged at 3,500 ×g for 10 min and re-suspended in RPMI 1640 medium at a concentration of 10⁶ cells/mL. Lactate and hydrogen peroxide were added to the RPMI broth with 8% MRS and adjusted to a concentration ranging from 0.5 to 1024 mM and 0.5 nM to 1024 mM, respectively. After 24 hours of incubation at 37°C, the turbidity of all well broth was visually observed, and the lowest concentration of lactate or hydrogen peroxide that suppressed the increased growth was determined as the MIC.

The efficacy of the anti-biofilm activities of lactobacilli was determined by adding the culture supernatant of Lactobacillus to the preformed biofilm. C. albicans HB-10, which formed a mature biofilm in the assay described above, was selected and used for subsequent inhibition assays. A mature biofilm of C. albicans HB-10 was formed in a 96-well microtiter plate after 24 h of incubation under the same conditions as the biofilm formation and viability assay. The planktonic cells were aspirated from each well and washed twice with PBS. Cell-free supernatant extracted from a single Lactobacillus strain was added to each well at a final concentration of 8% (v/v) and incubated for 24 h. Culture supernatants were aspirated from each well and washed twice with PBS. Biofilm formation was quantified using CV and WST-1, as described above. The metabolic activity of the residual biofilm after the exposure of the Lactobacillus culture supernatant was quantified using WST-1 as described above. Eight replicate wells were used for each strains, and experiments was repeated three times, independently.

According to the hyphal formation method in the RPMI broth described previously (Wang et al., 2017), we estimated the effect of hyphal formation inhibition of C. albicans yeast-to-hyphal transition in the presence of Lactobacillus culture supernatants. Lactobacilli with strong inhibition of biofilm formation and those with low inhibition were selected. C. albicans HB-10 cells from overnight culture were washed with PBS and re-suspended at approximately 10⁶ CFU/mL in RPMI 1640 medium buffered with MOPS. The yeast cell suspensions were then incubated with or without Lactobacillus culture supernatant at 37°C for 3 h. Quantification of the inhibitory effect of Lactobacillus on hyphal formation was performed using a light microscope (AxioCam MRc5; Carl Zeiss, Jena, Germany). The percentage of hyphal formation was calculated by obtaining the ratio of total number of C. albicans cells with hyphae to the total number of C. albicans cells counted. The number of yeast and hyphae (total cells) was counted using a hemocytometer (Supplementary Table 3).

Human cervical cancer HeLa cells (RCB0007; Riken BRC Cell Bank, similar to ATCC CCL2) were grown in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, MA, USA) supplemented with 10% (v/v) fetal bovine serum (Thermo Fisher Scientific, MA, USA) and 1% (v/v) penicillin and streptomycin (FUJIFULM Wako Pure Chemical, Co., Ltd., Osaka, Japan) at 37°C under 5% CO₂ and humidity. HeLa cells were seeded into a 12-well plate (AGC Techno Glass Co., Ltd., Shizuoka, Japan) at approximately 1.0×10⁵ cells per well and grown to confluence. After 90% confluency, each well was washed twice with PBS. C. albicans HB-10 and lactobacilli were grown under the conditions described above. Briefly, a single colony of C. albicans HB-10 and lactobacilli was inoculated into YPD and MRS broth, respectively. After 48 h of incubation, both C. albicans HB-10 and lactobacilli cells were collected using centrifugation at 3,500 ×g for 10 min and re-suspended in DMEM. The suspension of C. albicans HB-10 and lactobacilli contained approximately 1.0×10⁷ CFU/mL. To the HeLa cell culture well, 100 μL of 10-fold serial dilutions of lactobacilli suspensions was added and incubated at 37°C for 1 h under 5% CO₂. Subsequently, 100 μL of 10-fold serial dilutions of C. albicans HB-10 suspensions was added to each well and incubated for 1 h under the same conditions to allow C. albicans HB-10 to adhere to cells. After incubation, each well was washed twice with PBS to remove non-adherent C. albicans cells and then treated with 0.05% trypsin-EDTA (Nacalai Tesque, Inc., Kyoto, Japan). The inhibitory rate of adhesion was calculated as the number of C. albicans cells that adhered to HeLa cells with Lactobacillus pre-treatment, per the number of C. albicans cells that adhered to HeLa cells in the absence of lactobacilli.

Statistical analyses were performed using R and RStudio (versions 4.0.3 and 1.4.1106, respectively). Mann–Whitney U-test was used to determine significant differences between DMEM control and different lactobacilli. One-way analysis of variance was used to compare multiple groups. Statistical significance was set at p values <0.05. Correlations between growth and metabolites (lactate and hydrogen peroxide) were determined using Spearman’s rank correlation coefficient.

Biofilm formation by C. albicans clinical isolates was assessed using the WST-1 formazan dye (Figure 1A).

Among these 45 strains, representative strains that reproduced well and showed significant differences in the CV assay were selected as high and low biofilm-producing strains and renamed as C. albicans HB-1 and HB-10 and C. albicans LB-9 and LB-22, respectively (Figure 1B). The expression levels of the three genes, ECE1, HWP1, and YWP1, which regulate different stages of biofilm formation in C. albicans, were determined using qRT-PCR (Figure 1C).

Four C. albicans strains with gene expression of HWP1, ECE1, and YWP1 exhibited bar plots normalized by C. albicans HB-10 gene expression levels. These gene expression levels were calculated according to ACT1 gene expression levels. Similar to the phenotypic biofilm-forming analysis, the relative gene expression levels of ECE1 and HWP1 in the HB-10 strain were significantly higher than those in the LB-9 (10.53-fold and 3.21-fold, respectively) and LB-22 strains (31.21-fold and 8.72-fold, respectively) (p < 0.05). Interestingly, the HB-1 strain, which was a high biofilm producer and did not show such a large difference in biofilm-forming ability, had significantly lower expression levels of these genes than the HB-10 strain. In contrast, YWP1, which suppressed initial adhesion, was the lowest in the HB-10 strain.

Vaginal lactobacilli produce various metabolites that exhibit antifungal activity. Cell-free culture supernatants extracted from 27 strains of Lactobacillus belonging to five species after 48-h incubation were characterized (Figure 2).

The lactate and hydrogen peroxide production of lactobacilli used in this study ranged from 42.1 to 201.7 mM and 38.7 to 170.8 nM, respectively. There was no correlation between lactate production and hydrogen peroxide production (r = 0.426; p = 0.217). In contrast, OD corresponding to the growth of Lactobacilli and lactate levels at the 72-h sampling point showed moderately positive correlations (r = 0.667; p < 0.001), while OD and hydrogen peroxide levels showed no correlation (r = 0.265; p = 0.181) (Supplementary Figure 2). In L. jensenii, the average hydrogen peroxide production was the highest compared to other species (mean 146.0 nM), although lactate production was not as high (ranging from 81.3 to 126.5 mM).

In a typical experiment using 96 well plates, biofilm formation takes 24 h to reach confluency. We investigated the effects of Lactobacillus culture supernatants on preformed biofilms (Figure 3).

The addition of Lactobacillus culture supernatant resulted in 24.3%-91.8% relative WST-1 readings compared with those of non-added control. The culture supernatants of L. crispatus, L. gasseri, L. jensenii, and L. vaginalis with average relative WST-1 readings were approximately 52.5%, 43.1%, 57.2%, and 58.9%, respectively. The effect of different lactobacilli species on preformed biofilms of C. albicans HB-10 did not differ significantly. A moderate negative correlation was found between Lactobacillus lactate production and WST-1 readings (r = −0.625; p <0.001), but no correlation was observed between hydrogen peroxide production and WST-1 readings (Supplementary Figure 3).

We added several concentrations of the standards to the biofilm to reproduce lactate and hydrogen peroxide as metabolites in the culture supernatant (Supplementary Figure 4). The results showed that lactate concentrations had lower WST values than controls at all concentrations except 4 mM and a concentration-dependent effect on WST values (r = -0.930; p = 0.001). In contrast, hydrogen peroxide had no concentration-dependent effect on the preformed biofilm, with WST values not significantly different from the control at all concentrations. Lactate and hydrogen peroxide further showed no additive or synergistic effects on the preformed biofilms. A strong effect on the preformed biofilm was observed when the culture supernatant of the L. crispatus 35-1 strain was added (final concentration 14.1 mM), which was consistent with the WST-1 values for the biofilm when the 16 mM lactate standard was added (Figure 3 and Supplementary Figure 4).

The inhibitory effect of Lactobacillus spp. on C. albicans yeast cell growth was also evaluated. Significant growth inhibition was shown in 4/27 (14.8%) of the strains with the addition of culture supernatant of each Lactobacillus as follows: L. crispatus strain 23-1 and 20-2, L. gasseri strain 31-1 and L. rhamnosus strain 14-1. Interestingly, three of these strains showed lactate and hydrogen peroxide production above 165 mM and 120 nM, respectively (Supplementary Figure 7). The MICs for lactate and hydrogen peroxide standard for C. albicans HB-10 samples were 512 mM and 20 mM, respectively.

The effect on the rate of hyphal formation was compared for Lactobacillus culture supernatants that exhibited significant differences in their WST values to the preformed biofilm and for those that did not (Figure 4).

Yeast, hyphae, and pseudohyphae were identified under a microscope (Supplementary Figure 8). The addition of MRS medium control resulted in 54.95 ± 9.61% of the hyphae, and pseudohyphae were identified after 3 h of incubation. Compared to MRS control alone, lactate standards showed a concentration-dependent decrease in hyphal formation at 16 to 64 mM (p<0.05), while only 64 nM of hydrogen peroxide showed a significant difference. The percentage of hyphal formation by Lactobacillus culture supernatants ranged from 18.11 to 54.77%. In terms of the percentage of a hyphal formation relative to untreated MRS, significant decreases were observed with the addition of culture supernatant in L. crispatus strain 35-1 (32.97 ± 13.29%) and 53-1 (39.10 ± 7.40%), L. gasseri strain 45-3-1 (56.99 ± 6.90%) and 32-2 (70.75 ± 4.15%), and L. vaginalis strain 41-1 (70.60 ± 8.12%). All of these showed final lactate concentrations >50 mM. In contrast, although there were no differences in lactate and hydrogen peroxide metabolite profiles between L. gasseri strains 45-3-1 and 45-3-2, only strain 45-3-1 showed significantly lower hyphal formation than the MRS control.

The inhibition of C. albicans yeast adhesion to human epithelial cells by Lactobacillus bacterial cells was assessed (Figure 5).

L. crispatus strain 35-1 showed a more efficient inhibition of C. albicans HB-10 (adhesion rate: 68.29 ± 6.90%). In contrast, L. crispatus strain 3-1-2 showed no statistically significant difference (adhesion rate: 93.90 ± 25.87%). Interestingly, L. gasseri strain 45-3-1, which showed inhibition of hyphal formation, also showed a significant reduction in C. albicans HB-10 adhesion (80.95 ± 3.17%), whereas strain 45-3-2 showed no inhibitory effect on initial adhesion (89.68 ± 2.38%).