C. albicans strain SC5314 was obtained from the Department of Mycology, Institute of Dermatology, Chinese Academy of Medical Sciences. The C. albicans strain in a cryopreservation tube was removed from an -80°C freezer, scraped with an aseptic inoculation ring, and grown overnight in 4 ml of YPD medium (10 g per liter yeast extract, 20 g per liter peptone, 20 g per liter glucose) at 30°C. Then, 200 μL of the C. albicans solution was added to 4 ml of newly prepared YPD liquid medium. After mixing, the mixture was cultured in a shaker at 30°C and 180 rpm for 6 h. After washing twice with PBS, the cells were resuspended in keratinocyte serum-free medium (KSFM) (Gibco, Invitrogen, Carlsbad, CA, USA) and counted with a hemocytometer. The cell density was maintained at 5 × 10⁴ cells/ml for subsequent seeding.

Leuk-1, an immortalized human oral mucosal epithelial cell line, was a generous gift from Professor Li Mao in the Department of Oncology and Diagnostic Sciences, University of Maryland Dental School, Baltimore, MD, USA. The human oral mucosal epithelial cell line Leuk-1 was expanded and passaged in KSFM at 37°C in 5% CO₂, routinely grown to 80% confluency, and trypsinized with 0.25% trypsin-EDTA. Leuk-1 cells were Mycoplasma-tested to avoid Mycoplasma contamination.

EVs derived from Leuk-1 cells (Leuk-1-EVs) were purified by sequential ultracentrifugation (17) ( Figure 1A ). Briefly, Leuk-1 cells were cultured in KSFM medium for 48 h and then centrifuged for 10 minutes at 500 × g to remove cell contamination, 20 minutes at 3,000 × g to remove apoptotic bodies and large cell debris, and 20 minutes at 12,000 × g to remove large microvesicles. After that, EVs were collected by ultracentrifugation in 31 mL ultracentrifugation tubes (#355618, Beckman Coulter, USA) at 110,000 × g for 70 min, washed in PBS, and pelleted again by ultracentrifugation at 110,000 × g in 70 Ti fixed-angle rotors in a Beckman Coulter Optima XPN/XE ultracentrifuge. Finally, the EVs were resuspended in 300 μl of PBS and stored at −80°C.

The ultrastructure and size of EVs were analyzed by transmission electron microscopy (TEM) (Hitachi, HT7800/HT7700, Japan) and subjected to nanoparticle tracking analysis (NTA) with a Zeta View^® system (Particle Metrix, Germany). The protein concentration was measured by BCA assay (Enhanced BCA Protein Assay Kit, Beyotime Institute Biotechnology, China). Western blotting was performed to detect the protein markers of EVs, including CD63 (ab, 59479, Abcam, Cambridge, UK), TSG101 (ab 83, Abcam, Cambridge, UK), and HSP70 (10995-1-AP, Proteintech, Wuhan, China).

C. albicans and Leuk-1-EVs were prepared as described above. The experimental process is shown in Figure 1A .

GW4869 (dihydrochloride hydrate) is a cell-permeable but selective inhibitor of neutral sphingomyelinase (an essential enzyme required for EV production and release) (18). This experiment was performed to observe the growth of C. albicans after inhibition of Leuk-1 cell exosome secretion under an optical microscope.

The above experiments confirmed the important antifungal effect of Leuk-1-EVs by inhibiting EVs. To more fully prove the anti-C. albicans effect of oral mucosal epithelial cell EVs, we observed the antifungal effect of Leuk-1-EVs on C. albicans after coincubation with Leuk-1 cells.

Leuk-1 cells (2 × 10⁵) were seeded into 96-well plates ( Figure 1A ) and cultured for 24 h (37°C, 5% CO₂). The supernatant was then removed and discarded, and EVs were added to the C. albicans + EVs and C. albicans + Leuk-1 + EVs wells at a final concentration of 120 μg/ml. After culturing for 24 h, C. albicans suspensions were prepared. The concentrations of C. albicans cell suspensions added to Leuk-1 and the following steps were the same as those described above. The Leuk-1 cells in the wells were counted, and C. albicans was added to the C. albicans + EVs, C. albicans + Leuk-1 + EVs, C. albicans and C. albicans + Leuk-1 suspensions and mixed gently by pipetting. The method used to observe the changes in germ tubes, hyphae and microcolony formation was the same as that used above.

The experimental procedures were performed in accordance with the Guide for the Welfare and Use of Laboratory Animals and approved by the local Ethics Committee [IRB Approval Number: 2015NL-001 (KS) & 2018NL-008(KS)]. A mouse model of pseudomembranous oral candidiasis was used in order to characterize mucosal biofilms in vivo following the protocol (19, 20) ( Figure 1A ). In the experiments 6–8 week old female C57BL/6J mice were infected with strain SC5314. The mice were raised in cages and were free to get food and water. The photoperiods were adjusted to 12 h of light and 12 h of darkness, and the temperature was kept constant at 22°C. One day prior to C. albicans infection mice were immunosuppressed by subcutaneous injection with prednisolone (100mg/kg) (Xianju Pharmaceutical Co., Zhejiang, China). Following prednisolone injection, 0.8g/L tetracycline hydrochloride (Solarbio Science & Technology Co., Beijing, China) was administered in drinking water. To deliver C. albicans challenge mice were anaesthetized with sodium pentobarbital (50 mg/kg i.p.) and a cotton swab soaked with C. albicans cell suspension (5×108 yeast/ml) was used to swab the entire oral cavity. A cotton pellet soaked with C. albicans suspension was left under the tongue and was removed before the mice awoke.

The infected mice were randomly divided into the C. albicans group and C. albicans + EVs group (5 mice per group). The mice in the C. albicans group or C. albicans + EVs group received submucosal injection of PBS or Leuk-1-EVs in the tongue, respectively. Each mouse in the C. albicans + EVs group was injected with 20 μl Leuk-1-EVs (360 µg/ml). Each mouse in the C. albicans group was injected with 20 μl PBS. The procedures were performed once per day and repeated 3 days. At the end of experiment period, tongues were dissected, harvested and fixed after euthanasia of the mice.

The white lesions on the tongue of infected mice were evaluated by two observers who were blinded to the experimental data. The lesions were measured by scoring the pseudomembrane from 0 to 4 (20, 21), which were expressed as follows: 0, normal; 1, white lesion in less than 20%; 2, white lesion in less than 50% but more than 21%; 3, white lesion in less than 90% but more than 51%; 4, thick white lesion in more than 91%.

For hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS) stainings, tongues of mice were harvested and placed in 10% formalin for 24 hours. Paraffin embedded tongue specimens were cut into 5-μm-thick slices and stained with H&E and PAS, and then observed under a microscope. For SEM evaluations, tongue tissue specimens were fixed with 4% electron microscope fixative for 4 h, and dehydrated with a gradient of increasing alcohol concentrations. Then, they were dried using the critical point method. The samples were gold-sputtered (Hitachi, MC1000) to generate conductivity and tested with SEM (Hitachi, SU8100).

The severity of the inflammation was evaluated on a scale from 0 to 3 by two observers who were blinded to the experimental data depending on the extent of the inflammatory cell infiltration (grade 0, no more than 5 inflammatory cells in each high-power field; grade 1, small amount of inflammatory cells, limited to the superficial mucosal epithelium, no more than 1/3 of the mucosal epithelium; grade 2, moderate amount of inflammatory cells, no more than 2/3 of the mucosal epithelium; grade 3, large amount of inflammatory cells, occupation of the whole mucosal epithelium) (22).

The invasion degree of C. albicans was also evaluated on a scale from 0 to 4 by two observers who were blinded to the experimental data depending on the invasion range of C. albicans in the mucosal epithelium (grade 0, none; grade 1, infestation range from 1% to 25%; grade 2, infestation range from 26% to 50%; grade 3, infestation range from 51% to 75%; grade 4, infestation range from 76% to 100% (23).

All experiments were performed at least three times in duplicate. All images were analyzed using Image-Pro Plus image analysis software Version 7.0.1 (Media Cybernetics Inc). All statistical analyses were performed using the Statistical Package for the Social Sciences 22.0. The data were analyzed using the independent Student t-test or one-factor ANOVA if normally distributed, and the Mann-Whitney U test if not normally distributed. ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

EVs were isolated from Leuk-1 cell culture supernatant through ultracentrifugation ( Figure 1A ) and examined by NTA, TEM and Western blotting. NTA indicated a typical average EV size distribution starting at 120 nm ( Figure 1B ), similar to that of EVs released in other body fluid, and TEM verified that the collected vesicles were cup-shaped, rounded vesicles of 30–200 nm ( Figure 1C ). Western blotting analysis revealed that these vesicles were positive for CD63, HSP70 and TSG101 ( Figure 1D ), demonstrating that these vesicles were EVs. Therefore, we successfully isolated and characterized EVs derived from the human oral mucosal epithelial cells, which could be used to verify their antimicrobial effect.

To evaluate the effect of Leuk-1-EVs on C. albicans, we treated C. albicans with different concentrations of Leuk-1-EVs. Compared with the other groups, the groups treated with higher concentrations of EVs (120 µg/ml and 360 µg/ml) exhibited significantly decreased hyphae formation at 12 h and microcolony formation at 24 h and 48 h ( Figures 2B–D ). The 360 µg/ml group showed a more obvious antifungal effect than the 120 µg/ml group, while no significant changes in germ tube formation were observed at 2 h ( Figure 2A ). Leuk-1-EVs inhibited C. albicans growth in a concentration-dependent manner. These results indicated that Leuk-1-EVs could exert antifungal effects by inhibiting C. albicans growth and hyphae formation.

Confocal microscopy revealed the presence of irregular protrusions on the hyphal wall after treatment with high concentrations of EVs (120 μg/ml and 360 μg/ml) ( Figure 3 ). The white arrows in Figure 3A denote the protrusions on the hyphal wall. The results showed that there were significantly more protrusions on hyphae in the high concentration treatment groups than those in the low concentration treatment group and control group (P < 0.001) ( Figure 3B ). To further investigate, we used SEM and TEM to examine changes in the micromorphology and ultrastructure of C. albicans. In SEM, the untreated C. albicans wall was even and smooth ( Figure 4A ). In the experimental group, more spores were found, and there were more obvious protrusions on hyphae ( Figure 4B ). The results also showed that C. albicans in the experimental group had more spores and hyphal protrusions than that in the control group (P < 0.001) ( Figures 4C, D ). These specific changes are indicated by black arrows in Figure 4B . In TEM, the untreated C. albicans cells displayed smooth and intact cell walls, homogeneous cytoplasm and intact organelles ( Figure 5A ). In contrast, exposure to EVs caused some alterations, including uneven plasma density (green arrows in Figure 5B ), increased mucus protein (black arrows in Figure 5B ) and protrusions (brown arrows in Figure 5B ) on the cell wall, incomplete cell walls (red arrows in Figure 5B ), an inner roll of the cell membrane (yellow arrows in Figure 5B ), evidently swollen mitochondria, organelle swelling, and uneven cytoplasm (purple arrows in Figure 5B ). We also analyzed the average density and thickness of the cell wall of C. albicans in the two groups. The results showed that the mean IOD of the cell wall in the experimental group (4.37 ± 0.33) was significantly lower than that in the control group (10.09 ± 0.51) (P < 0.001) ( Figure 5C ), and the mean thickness of cell wall in the experimental group (230.76 ± 11.35 nm) was significantly higher than that in the control group (99.28 ± 7.94 nm) (P < 0.001) ( Figure 5D ). These results further suggested that Leuk-1-EVs could play antifungal effects by inducing damages of cell wall integrity and organelle structure.

GW4869 is a specific inhibitor of nSMases (also known as SMPD2 and 3) and has been shown to inhibit exosome release from cells (8). To screen the optimal concentration of GW4869, the viabilities of Leuk-1 cells treated with varying concentrations of GW4869 were measured by CCK-8 assay ( Figure 6A ). The cells treated with high concentrations of GW4869 (10 and 20 µM) exhibited significantly lower viability (81.44% and 56.70%, respectively) than the cells in all of the other groups ( Figure 6A ). Therefore, the optimal concentration was set as 7.5 µM. To investigate the effect of EV release decrease on anti-candida activity of Leuk-1 cells, we evaluated the formation of germ tubes, hyphae, and microcolonies of C. albicans in the cocultures after treatment of Leuk-1 cells with GW4869.

The formation of germ tubes, hyphae (12 h), and colonies (24 h and 48 h) of C. albicans SC5314 was observed at 2 h after the suppression of Leuk-1 cell exosome secretion through a microscope, as shown in Figure 6 . The results showed that there was no difference in the rate of germ tube formation between the C. albicans + Leuk-1 and C. albicans + Leuk-1+ GW4869 groups (P > 0.05). Compared with the control group, the germ tube formation rates of the C. albicans + Leuk-1 group and the C. albicans + Leuk-1+ GW4869 group were significantly lower (P < 0.001) ( Figure 6B ). The 12 h, 24 h, and 48 h optical density analysis results are shown in Figures 6C–E , respectively. The results show that the optical densities of hyphae and microcolonies in the C. albicans + Leuk-1 + GW4869 group were higher than those in the C. albicans + Leuk-1 group (P<0.001, 12h; P<0.01, 24h; P<0.01, 48h, respectively) and that the optical densities of hyphae and microcolonies in the C. albicans + Leuk-1 group and the C. albicans + Leuk-1+ GW4869 group were lower than those in the C. albicans group (P < 0.001). It was shown that reduced EVs release could impair antifungal effects of Leuk-1 cells on C. albicans. Therefore, these results indicated that Leuk-1 could play antifungal effects by releasing EVs.

The formation of germ tubes, hyphae (12 h), and microcolonies (24 h and 48 h) of C. albicans under the synergistic effects of Leuk-1 cells and Leuk-1-EVs was observed under a microscope, as shown in Figure 7 . The results showed that the rate of germ tube formation in the C. albicans + EVs group was not significantly different from that in the C. albicans group at 2 h (P > 0.05). The germ tube formation rates of the C. albicans + Leuk-1 group and the C. albicans + Leuk-1 + EVs group were significantly lower than that of the C. albicans group (P < 0.001) ( Figure 7A ). The 12h, 24h, and 48 h optical density analysis results are shown in Figures 7B–D , respectively.

The hyphal and microcolony optical densities of the C. albicans + Leuk-1 group and the C. albicans + Leuk-1+ EVs group were lower than those of the C. albicans group (P < 0.001), and the hyphal and microcolony density of the C. albicans + Leuk-1 + EVs group was lower than those of the C. albicans + EVs group and the C. albicans + Leuk-1 group (P < 0.01, 12h; P < 0.001, 24h; P < 0.001, 48h, respectively). At 12 h and 24 h, the hyphal and microcolony density of the C. albicans + EVs group was lower than those of the C. albicans + Leuk-1 group (P < 0.001, 12h; P<0.01, 24h, respectively); at 48 h, there was no significant difference in the microcolony density between the C. albicans + EVs group and C. albicans + Leuk-1 group (P > 0.05). These results further suggested that Leuk-1-EVs could increase antifungal effects of Leuk-1 cells by suppressing C. albicans growth synergistically.

The mouse model of oral candidiasis was successfully constructed. The white lesions on the tongue consisting of white patches or pseudomembranes were observed in the C. albicans group. There were smaller white lesions in the C. albicans + EVs group compared with the C. albicans group ( Figure 8A ). The score of white lesions in the C. albicans + EVs group (1.20 ± 0.45) significantly decreased compared with that in the C. albicans group (3.00 ± 0.71) (P < 0.01) ( Figure 8B ). These results indicated that Leuk-1-EVs treatment could exert antifungal effects in vivo by alleviating C. albicans infection of oral mucosa.

The histopathological findings of oral mucosal tissues were shown in Figure 9 . The results of H&E staining showed that normal filamentous papillae could be observed in the C. albicans + EVs group. Instead, in the C. albicans group filamentous papillae was absent, but thickened cornified layer and microabscess could be noted in the epithelium ( Figure 9A ). There were significantly fewer inflammatory cells in the C. albicans + EVs group compared with the C. albicans group ( Figure 9A ). The intensity of inflammatory cell infiltration in the C. albicans + EVs group (0.33 ± 0.23) was significantly lower than that in the C. albicans group (2.67 ± 0.24) (P < 0.001) ( Figure 9B ). These results suggested that Leuk-1-EVs treatment could decrease inflammatory response of oral mucosa to C. albicans infection.

The findings of PAS staining were shown in Figure 9 . Our results showed that there were more hyphae of C. albicans in the tongue mucosa in the C. albicans group than those in the C. albicans + EVs group ( Figure 9C ). Clearly, C. albicans hyphae could be observed in the mucosal epithelium of the C. albicans group ( Figure 9C ). Filamentous papillae disappeared in the C. albicans group ( Figure 9C ). Compared with the C. albicans group (2.33 ± 0.24), the invasion degree of hyphae significantly reduced in the C. albicans + EVs group (0.53 ± 0.45) (P < 0.001) ( Figure 9D ).

The SEM findings were shown in Figure 10 . Consistent with the PAS staining results, the SEM results showed that there were more hyphae of C. albicans on the dorsum of the tongue in the C. albicans group than those in the C. albicans + EVs group ( Figures 10A, B ). Biofilm and keratin desquamation could be commonly observed in the C. albicans group. Analysis of hyphae and spores number also showed that the C. albicans group had significantly more hyphae and spores than the C. albicans + EVs group (8.80 ± 0.84 vs 2.20 ± 0.56, P < 0.001; 15.73 ± 0.64 vs 4.93 ± 0.64, P < 0.001, respectively) ( Figures 10C, D ). Consequently, our results indicated that Leuk-1-EVs treatment could play antifungal effects by limiting hyphal invasion of C. albicans into the oral mucosa.