Saccharomyces cerevisiae reduces vulvovaginal candidiasis severity through modulation of fungal pathogenicity and inflammatory responses Authors: Mart Sillen, Dania El Abyad, Nina Vreys, Silke Baldewijns, Ilse Palmans, Odessa Van Goethem, Kar On Cheng, Greetje Vande Velde, Dolly E. Montaño, Nele Berghmans, Mieke Gouwy, Agustin R. Sharpe, Lies De Groef, Axel Dietschmann, Paul Vandecruys, Sarah Lebeer, Liesbeth Demuyser, Mark S. Gresnigt, Patrick Van Dijck Published: 2026 ABSTRACT Vulvovaginal candidiasis (VVC) affects millions of women globally and is characterized by multifactorial immunopathology, with Candida albicans virulence driving disease progression through epithelial tissue damage and neutrophil hyperactivation and dysfunction contributing to disease severity. Here, we explored the multifaceted nature of S. cerevisiae as a live-biotherapeutic to attenuate C. albicans virulence and modulate host immune responses during VVC. We identified an S. cerevisiae isolate (Sc3458) that targets multiple aspects of C. albicans virulence, including fungal proliferation, adhesion, and hyphal morphogenesis, collectively impairing biofilm formation and disrupting pathogenic potential. These effects were linked to transcriptional reprogramming in C. albicans, marked by metabolic stress and downregulation of virulence- and biofilm-related genes. Additionally, S. cerevisiae (Sc3458) reduced inflammatory responses and neutrophil hyperactivation, while preserving neutrophil antimicrobial functionality. Altogether, this translated to an improved control of infection and dampened VVC-associated hyperinflammation in a murine VVC model. These findings position S. cerevisiae strain Sc3458 as a promising candidate live biotherapeutic for VVC treatment. Further clinical validation in human cohorts is essential to confirm its therapeutic efficacy and to assess how Sc3458 performs relative to established market-standard formulations. Introduction Vulvovaginal candidiasis (VVC), an infection of the vaginal mucosa caused by yeast of the genus Candida, is the second most prevalent form of vaginitis, posing a significant health concern1,2. VVC affects up to 75% of women at least once, regardless of their overall health status3. Moreover, approximately 7–9% of women endure recurrent vulvovaginal candidiasis (RVVC), defined by three or more infections annually4,5, contributing to an estimated global burden of 138 million cases per year6. (R)VVC manifests distressing symptoms, such as white clumpy discharge, burning, redness, itching in the vulva and vagina, and dyspareunia, significantly impacting mental well-being3,7. Candida albicans, the most prevalent causative species8,9, induces VVC through hyperinflammation, driven by Candida virulence and maintained by the innate immune system, resulting in compromised neutrophil-mediated fungal clearance8,10. VVC manifests when the fungal burden exceeds an epithelial threshold, triggering epithelial pro-inflammatory signaling cascades that result in the release of inflammatory mediators8. This process catalyzes downstream inflammatory responses that facilitate the recruitment of immune cells, primarily neutrophils11. However, the hyperinflammation and specific characteristics of the vaginal niche impair neutrophils to effectively deploy their antifungal mechanisms, leading to tissue damage and perpetuating a cycle of inflammation and exacerbation of symptoms11. Being a multifaceted disease, VVC presents multiple therapeutic opportunities, including antifungal drugs and vaccines still under research3,12. While antifungals are central to RVVC management, prolonged use often fosters resistance4,5. Coupled with the complex nature of VVC, this leaves many patients insufficiently treated. Moreover, antifungals only indirectly impact inflammatory responses and fail to alleviate symptoms when non-self-limiting inflammation drives disease progression13. Therefore, alternative treatment strategies that address the multifactorial nature of VVC are urgently needed. Given the essential role of vaginal microbial communities in maintaining vaginal health, live microbial-biotherapeutic products (LBPs) present a promising avenue14,15. These are defined as live microorganisms intended for the prevention or treatment of disease16, and several Lactobacillaceae-based LBPs have shown efficacy in experimental and clinical VVC17,18. However, as infections are highly patient-specific, treatment efficacy is not universal19. Although yeasts are generally not dominant members of the vaginal microbiome, patients with RVVC, compared to healthy individuals, exhibit reduced levels of colonization with Saccharomyces spp., making these yeasts intriguing candidates for further study20. S. cerevisiae especially shows promise for use in patients undergoing antibiotic therapy, a common trigger for VVC development2. Recognized for its benefits in gastrointestinal disorders21,22, one S. cerevisiae strain has also shown efficacy in the context of VVC, demonstrating the ability to co-aggregate with C. albicans, inhibit its adherence to epithelial cells, suppress key virulence factors, and protect epithelial cells from Candida-induced damage23,24. Its effects likely arise from mechanisms such as pathogen binding, production of antimicrobial peptides, trophic effects, and immune-modulatory properties22,25. In this study, we focused on the potential of S. cerevisiae in counteracting C. albicans-induced VVC in lab-based and murine models. We conducted a comprehensive screening of a broad S. cerevisiae isolate library to assess their potential as therapeutic microbes against VVC. We explored how S. cerevisiae antagonizes C. albicans pathogenicity and modulates immune responses, focusing on insights gained from multispecies transcriptomics combined with in vitro infection models. To validate these findings in a physiologically relevant context, we treated mice with S. cerevisiae in the context of a well-established VVC model. Through examining the tripartite interactions between C. albicans, vaginal epithelial cells, and neutrophils, both in vitro and in vivo, we uncovered mechanistic insights into how S. cerevisiae acts as a live biotherapeutic for VVC. C. albicans virulence inhibition by physical constraint, metabolic interference, and transcriptional reprogramming resulted in a dampened pro-inflammatory milieu and enhanced neutrophil-mediated fungal clearance, two critical determinants dysregulated during symptomatic VVC. These findings establish a robust preclinical foundation for the translational development of S. cerevisiae-based interventions in human VVC. Results S. cerevisiae strain-specific inhibition of C. albicans Given large strain-to-strain variability within microbial species19,26, a screening was performed to identify S. cerevisiae strains with inhibitory properties against C. albicans, a key driver of VVC pathology10. We screened 70 diverse S. cerevisiae isolates, including food, lab strains, probiotics, beverages, and human sources (Table S1). Using assays for C. albicans CA3153 growth inhibition in vagina-simulative medium (VSM) and RPMI 1640 (RPMI), adhesion to A-431 vaginal epithelial cells (VECs), and filamentation inhibition, several effective S. cerevisiae strains were identified (Fig. S1). Notably, inhibitory capacity was not linked to the strain’s origin. From the 70 strains initially screened, the most promising candidates were shortlisted based on top performance in inhibiting C. albicans virulence, with those showing consistent efficacy across multiple assays prioritized. Based on these criteria, strains 4609, 6883, 6886, 6888, 6894, 3458, 3471, 3076, and Y.00240 were selected (Table S1). Two additional control strains were also included: the lab strain S288c and the probiotic standard CNCM I-3856, currently marketed for gastrointestinal (ibSium®) and vaginal (Quatreflora®) health24. S. cerevisiae reduces C. albicans pathogenicity on multiple levels To investigate the inhibitory potential of S. cerevisiae against C. albicans virulence in depth, we built on our extensive screening and focused on early infection processes across three C. albicans isolates SC5314, CA3153, and I49. Growth of C. albicans isolates in both VSM (Fig. 1A, B) and RPMI (Fig. S2A) was broadly inhibited by the selected S. cerevisiae strains, reducing C. albicans growth by up to 25–50% within 6 h and reaching up to 75% inhibition by 24 h. Notably, specific strains—C(6886), F(3458), G(3471), and IC(CNCM I-3856)—achieved higher levels of inhibition at earlier time points. Moreover, the growth inhibition was robust across all tested C. albicans isolates.Fig. 1S. cerevisiae isolates inhibit C. albicans growth, adhesion, and filamentation.A Inhibition of C. albicans (strains SC5314, CA3153, I49) growth by S. cerevisiae isolates (A–I), laboratory control S288c (LC), and probiotic standard CNCM I-3856 (IC) after 6 h (upper panel) and 24 h (lower panel) in VSM (n = 3). B Representative images of fluorescent C. albicans CA3153 co-cultured with S. cerevisiae isolates in VSM after 3 h. Images display growth and filamentation inhibition, with scale bars set at 100 µm (insets, 25 µm). C Inhibition of C. albicans adhesion to VECs by S. cerevisiae. The upper panel shows competition for initial adhesion (n = 3), while the lower panel shows displacement of already adhered C. albicans (SC5314, CA3153 n = 3; I-49 n = 4). D Inhibition of C. albicans filamentation in VSM by S. cerevisiae isolates after 3 h (n = 3). Beige hyphal and yeast-shaped cells represent C. albicans, purple cells represent S. cerevisiae, and pink cells represent VECs. (A, C, D). All bars represent mean ± standard deviation (SD), with individual data points representing the mean of technical replicates in the individual experiments. Data were transformed using a Box-Cox transformation to meet linear regression assumptions. Statistical significance was determined using a two-sided one-sample t-test against a theoretical mean of 0%, representing C. albicans alone. Significant differences are indicated as P-values. Schematic overview in panels (A–D) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. S. cerevisiae isolates inhibit C. albicans growth, adhesion, and filamentation. A Inhibition of C. albicans (strains SC5314, CA3153, I49) growth by S. cerevisiae isolates (A–I), laboratory control S288c (LC), and probiotic standard CNCM I-3856 (IC) after 6 h (upper panel) and 24 h (lower panel) in VSM (n = 3). B Representative images of fluorescent C. albicans CA3153 co-cultured with S. cerevisiae isolates in VSM after 3 h. Images display growth and filamentation inhibition, with scale bars set at 100 µm (insets, 25 µm). C Inhibition of C. albicans adhesion to VECs by S. cerevisiae. The upper panel shows competition for initial adhesion (n = 3), while the lower panel shows displacement of already adhered C. albicans (SC5314, CA3153 n = 3; I-49 n = 4). D Inhibition of C. albicans filamentation in VSM by S. cerevisiae isolates after 3 h (n = 3). Beige hyphal and yeast-shaped cells represent C. albicans, purple cells represent S. cerevisiae, and pink cells represent VECs. (A, C, D). All bars represent mean ± standard deviation (SD), with individual data points representing the mean of technical replicates in the individual experiments. Data were transformed using a Box-Cox transformation to meet linear regression assumptions. Statistical significance was determined using a two-sided one-sample t-test against a theoretical mean of 0%, representing C. albicans alone. Significant differences are indicated as P-values. Schematic overview in panels (A–D) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. C. albicans adhesion to VECs is crucial for the initiation of infection, invading the epithelial cells, and causing damage27. Simultaneous introduction of S. cerevisiae reduced C. albicans adhesion to VECs by ~30%, while some strains displaced up to 50% of pre-adhered C. albicans cells (Fig. 1C), disrupting established C. albicans–VEC interactions. Despite variability between S. cerevisiae strains and individual repetitions, certain strains consistently inhibited adherence of all three C. albicans isolates. Filamentous growth of C. albicans strengthens its adhesion and facilitates tissue invasion and damage of epithelial cells27. This further prompts C. albicans recognition and initiation of epithelial inflammatory responses8,28. S. cerevisiae strains (A(4609), C(6886), F(3458), G(3471), and H(3076)) inhibited (30–50%) or delayed hyphal growth in VSM after 3 h (Fig. 1D), though the effects were less pronounced in RPMI which is a stronger stimulus for filamentation (Fig. S2B). While all S. cerevisiae strains reduced C. albicans growth in VSM (Fig. 1A, B), not all inhibited the morphological transition, and some, B(6883), I(Y.00240), and LC(S288c), even promoted hyphal growth. S. cerevisiae mitigates C. albicans-induced epithelial damage and inflammation The inhibitory effect of S. cerevisiae on C. albicans growth, filamentation, and adhesion to VECs could imply decreased epithelial tissue damage and inflammatory responses28. S. cerevisiae co-cultured reduced epithelial damage inflicted by C. albicans, quantified by the release of the host cytoplasmic enzyme LDH (Fig. 2A, B). While certain S. cerevisiae strains—D(6888), E(6894), F(3458), G(3471)—markedly reduced cell damage, some—I(Y.00240), LC(S288c), and IC(CNCM I-3856)—had minimal effect or even exacerbated toxicity.Fig. 2S. cerevisiae isolates modulate epithelial activation and cluster into performance-based groups.A Schematic showing C. albicans-induced LDH release from damaged epithelial cells, triggering macrophage activation via IL-1α and neutrophil stimulation via IL-8 and GM-CSF. The diagram highlights immune signaling pathways involved in epithelial damage and C. albicans-driven immune cell stimulation. Beige hyphal cells represent C. albicans and pink cells represent VECs. B The effect of S. cerevisiae isolates (A-I), laboratory control (LC, S288c), and probiotic standard (IC, CNCM I-3856) on C. albicans-induced LDH release by VECs. Data show modulation of LDH levels across C. albicans strains SC5314, CA3153, and I49. C The effect of S. cerevisiae isolates on C. albicans-induced IL-1α release by A-431 VECs. D The effect of S. cerevisiae isolates on C. albicans-induced IL-8 release by VECs. E The effect of S. cerevisiae isolates on C. albicans-induced GM-CSF release by VECs. F Heatmap ranking S. cerevisiae isolates based on their performance across C. albicans virulence inhibition assays. Rankings are presented numerically, with lower numbers indicating stronger inhibitory performance. Two clusters were identified using k-means clustering (k = 2), as shown by the dendrogram on the left. G K-means cluster analysis (k = 2) of S. cerevisiae isolates based on their rankings in C. albicans virulence assays. Principal component analysis (PCA) shows two distinct clusters (Cluster 1, light blue; Cluster 2, dark blue), with PC1 (50.7%) and PC2 (17.4%) explaining the variance in isolate performance. Heatmap data represent the mean from n ≥ 2 independent experiments. Schematic overview in panel (A) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. S. cerevisiae isolates modulate epithelial activation and cluster into performance-based groups. A Schematic showing C. albicans-induced LDH release from damaged epithelial cells, triggering macrophage activation via IL-1α and neutrophil stimulation via IL-8 and GM-CSF. The diagram highlights immune signaling pathways involved in epithelial damage and C. albicans-driven immune cell stimulation. Beige hyphal cells represent C. albicans and pink cells represent VECs. B The effect of S. cerevisiae isolates (A-I), laboratory control (LC, S288c), and probiotic standard (IC, CNCM I-3856) on C. albicans-induced LDH release by VECs. Data show modulation of LDH levels across C. albicans strains SC5314, CA3153, and I49. C The effect of S. cerevisiae isolates on C. albicans-induced IL-1α release by A-431 VECs. D The effect of S. cerevisiae isolates on C. albicans-induced IL-8 release by VECs. E The effect of S. cerevisiae isolates on C. albicans-induced GM-CSF release by VECs. F Heatmap ranking S. cerevisiae isolates based on their performance across C. albicans virulence inhibition assays. Rankings are presented numerically, with lower numbers indicating stronger inhibitory performance. Two clusters were identified using k-means clustering (k = 2), as shown by the dendrogram on the left. G K-means cluster analysis (k = 2) of S. cerevisiae isolates based on their rankings in C. albicans virulence assays. Principal component analysis (PCA) shows two distinct clusters (Cluster 1, light blue; Cluster 2, dark blue), with PC1 (50.7%) and PC2 (17.4%) explaining the variance in isolate performance. Heatmap data represent the mean from n ≥ 2 independent experiments. Schematic overview in panel (A) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. Through the release of alarmins such as IL-1α, epithelial tissue damage can catalyze a pro-inflammatory state, activating neutrophils8. C. albicans infection alone triggered minimal IL-1α release (Fig. 2A, C), which was not significantly altered by protective S. cerevisiae strains, suggesting that additional factors may regulate its production. MAPK pathways, activated by excessive fungal burden and hyphae, exacerbate VVC pathology by recruiting and activating neutrophils11,28,29. S. cerevisiae strains that dampened fungal growth and filamentation also reduced IL-8 and GM-CSF responses of VECs (Fig. 2D, E), cytokines that enhance neutrophil activation, survival, and effector function (Fig. 1C). Importantly, the probiotic S. cerevisiae strain IC (CNCM I-3856) enhanced pro-inflammatory cytokine responses, amplifying signaling beyond the response to C. albicans. Unlike other S. cerevisiae strains, it also independently triggered pro-inflammatory responses in VECs in the absence of C. albicans (Fig. S3A). Cluster analysis reveals superior S. cerevisiae strains for VVC treatment Our data demonstrate that S. cerevisiae strains variably attenuate C. albicans virulence and alleviate the epithelial immune responses. We aimed to identify a strain that consistently exhibited superior performance across all assays. Multivariate cluster analysis using the collective data from previous assays revealed two distinct clusters (k-means, k = 2), explaining over 70% of the variance observed in the dataset (Fig. 2F, G). This significant level of explained variance indicates a strong association between the identified clusters and the observed performance outcomes. The first cluster, comprised of the S. cerevisiae strains that demonstrated superior efficacy in reducing C. albicans virulence and modulating immune responses, included A(4609), B(6883), C(6886), E(6894), F(3458), G(3471), and H(3076), suggesting a robust live biotherapeutic potential. The second cluster comprised of strains that exhibited overall low performance: D(6888), I(Y.00240), LC(S288c), and IC(CNCM I-3856). Based on the consistently high performance across all readouts, S. cerevisiae strain (F) 3458 (Sc3458) was selected for further investigation. Importantly, strain Sc3458 displayed a favorable safety profile relevant for LBP development. Sc3458, in contrast to the IC strain, did not induce epithelial inflammation (Fig. S3A). Neither did Sc3458 induce epithelial damage (Fig. S3B), nor formed biofilms on VECs (Fig. S3C). Moreover, β-hemolytic activity was absent for both Sc3458 and the IC strain (Fig. S3D), and both strains demonstrated tolerance to the acidic conditions of the vaginal niche (Fig. S3E). Collectively, these findings identify Sc3458 as a safe, robust, and promising candidate for further investigation as an LBP strategy for VVC. S. cerevisiae drives transcriptional reprogramming in C. albicans and VECs Multispecies transcriptomics was performed to identify the underlying mechanisms by which Sc3458 attenuates C. albicans virulence and modulates epithelial inflammatory responses. Principal component analysis (PCA) revealed distinct C. albicans gene expression at 24-h post-infection (hpi) with slight overlap at 6 hpi, depending on the presence of Sc3458 (Fig. 3A). Similarly, VECs showed distinct clusters in the presence of Sc3458 at 24 hpi (Fig. 3B). Despite the proximity of PCA clusters, particularly for C. albicans at 24 hpi, significantly differentially expressed genes (DEGs) were identified (Fig. 3C, upper panel). PCA highlights major variability in gene expression between conditions, but overlapping clusters can still harbor substantial transcriptional differences. At 6 hpi, C. albicans showed 9 DEGs (Fig. S4), but this increased to 211 DEGs at 24 hpi in response to Sc3458. For VECs, the impact of S. cerevisiae was more pronounced at 24 hpi, with 1848 DEGs observed (Fig. 3C, lower panel). Together, the DEG counts and PCA results underscore that Sc3458 drives significant reprogramming of gene expression in both C. albicans and VEC, particularly at 24 hpi.Fig. 3S. cerevisiae strain 3458 modulates C. albicans and epithelial cell transcriptional responses.A, B PCA of transcriptional profiles of C. albicans (Ca, A) and vaginal epithelial cells (VECs, B) at 6 and 24 hpi. The clustering of transcriptional profiles shows the impact of Sc3458 at both time points. C Number of DEGs in C. albicans and VECs at 6 and 24 hpi. The circle size reflects DEG counts, with a notable increase observed at 24 hpi. Genes were selected based on a Log2 fold change > 1 or <−1 and P-value < 0.05. D Volcano plot depicting significantly upregulated (red) and downregulated (blue) genes in C. albicans at 24 hpi in response to treatment with Sc3458. Genes were selected based on a Log2 fold change > 1 or <−1 and P-value < 0.05. Differential expression analysis (C, D) was performed using two-sided Wald tests with Benjamini–Hochberg false discovery rate (FDR) correction for multiple comparisons (adjusted P < 0.05). E, F Gene Ontology (GO) enrichment analysis of downregulated (D, blue) and upregulated (E, red) genes in C. albicans at 24 hpi. GO terms were identified using ShinyGO (http://bioinformatics.sdstate.edu/go/) and refined via Revigo (similarity threshold of 0.5, http://revigo.irb.hr/). GO enrichment analysis (E, F) was conducted using a two-sided hypergeometric test with Bonferroni correction for multiple testing (adjusted P < 0.05). Biological processes are ranked by –Log10(P-value) (Bonferroni corrected), and fold enrichment is visualized by a color gradient. Key processes include biofilm formation, cellular interactions, and metabolic regulation. Data are derived from n = 3 biological replicates. S. cerevisiae strain 3458 modulates C. albicans and epithelial cell transcriptional responses. A, B PCA of transcriptional profiles of C. albicans (Ca, A) and vaginal epithelial cells (VECs, B) at 6 and 24 hpi. The clustering of transcriptional profiles shows the impact of Sc3458 at both time points. C Number of DEGs in C. albicans and VECs at 6 and 24 hpi. The circle size reflects DEG counts, with a notable increase observed at 24 hpi. Genes were selected based on a Log2 fold change > 1 or <−1 and P-value < 0.05. D Volcano plot depicting significantly upregulated (red) and downregulated (blue) genes in C. albicans at 24 hpi in response to treatment with Sc3458. Genes were selected based on a Log2 fold change > 1 or <−1 and P-value < 0.05. Differential expression analysis (C, D) was performed using two-sided Wald tests with Benjamini–Hochberg false discovery rate (FDR) correction for multiple comparisons (adjusted P < 0.05). E, F Gene Ontology (GO) enrichment analysis of downregulated (D, blue) and upregulated (E, red) genes in C. albicans at 24 hpi. GO terms were identified using ShinyGO (http://bioinformatics.sdstate.edu/go/) and refined via Revigo (similarity threshold of 0.5, http://revigo.irb.hr/). GO enrichment analysis (E, F) was conducted using a two-sided hypergeometric test with Bonferroni correction for multiple testing (adjusted P < 0.05). Biological processes are ranked by –Log10(P-value) (Bonferroni corrected), and fold enrichment is visualized by a color gradient. Key processes include biofilm formation, cellular interactions, and metabolic regulation. Data are derived from n = 3 biological replicates. More specifically, a total of 131 genes were downregulated, while 80 genes were upregulated in C. albicans when S. cerevisiae was present during infection (Fig. 3D). Gene Ontology (GO) enrichment analysis revealed upregulation of “amino acid transport and metabolism” (AGP2, ALP1, SER2) and “sulfur compound biosynthesis” (SAM2, MET16, MET15, CYS3), reflecting metabolic shifts to counter nutrient competition and oxidative stress (Fig. 3F). Increased expression of ion transport genes (FRE10, CFL2, CFL5) suggests efforts to maintain ionic and redox balance under oxidative stress. Upregulated biosynthetic (ACS2, ACC1) and ribosomal genes (GAR1, RPL39) indicate heightened protein synthesis demands to manage stress and support growth. These stress-induced adaptations highlight that C. albicans attempts to reprogram its metabolism and homeostatic mechanisms to survive the antagonistic effects imposed by Sc3458 during infection. The analysis of downregulated genes revealed links to processes essential for C. albicans biofilm development and cell-cell interaction (Fig. 3E). GO term enrichment identified significant downregulation of genes involved in “single-species biofilm development” (both submerged and on inanimate substrates), “cell aggregation,” and “intraspecies interactions.” Key metabolic pathways for nutrient acquisition within the biofilm micro-niche were impacted, with genes involved in monosaccharide, hexose, and carbohydrate transport (HGT13, GAL1, MAL31) downregulated, suggesting compromised biofilm integrity and cellular function. Genes related to adhesion (TRY5, SUC1, ZCF28, SAP2/4/8, DSE1), morphogenesis (HGC1, YAK1, CSA2, MAL2), and biofilm maturation (SAP8, PGA6) were also suppressed, proposing significant disruption of biofilm structure and functionality. Direct interaction and metabolic activity of S. cerevisiae are essential for anti-biofilm effects on C. albicans C. albicans biofilm formation on epithelial cells during VVC exceeds the epithelial threshold, triggering inflammatory responses and hindering neutrophil-mediated fungal clearance29,30. While metabolic changes in C. albicans induced by probiotics are known to reduce pathogenicity31, their specific effects on biofilm formation remain unclear. As biofilms play a key role in VVC pathogenesis, we investigated how Sc3458 impacts C. albicans biofilm formation. Corresponding with reduced C. albicans growth and filamentation on VECs (Fig. 1), a broad downregulation of C. albicans biofilm-associated genes in response to the presence of Sc3458 is observed (Fig. 4A). Accordingly, microscopy proved that Sc3458 significantly reduced C. albicans biofilm biomass on VECs during both early (24 hpi) and mature biofilm (48 hpi) stages (Fig. 4B, C). Additionally, biofilm thickness was diminished by Sc3458 (Fig. 4B, D). These inhibitory effects were comparable to those observed with the industrial control CNCM I-3856. We hypothesized that S. cerevisiae might redistribute C. albicans cells between the biofilm associated with VECs and cells in the supernatant, given the known aggregation tendencies of both species32. Interestingly, redistribution of cells from the biofilm to the supernatant was observed upon Sc3458 treatment (Fig. 4E). A similar redistribution pattern was observed for IC (Fig. 4E), suggesting a conserved interference mechanism. While Sc3458 alone primarily remained in the supernatant, indicating weak adhesion to VECs (Fig. 4F), the presence of C. albicans resulted in its incorporation into the biofilm. Strikingly, Sc3458 treatment displaced C. albicans into the supernatant while embedding itself within the C. albicans biofilm, resulting in fewer C. albicans cells being directly in contact with VECs and capable of inflicting damage and eliciting inflammatory responses.Fig. 4S. cerevisiae reduces C. albicans biofilm biomass, alters gene expression, and disrupts biofilm integrity.A Heatmap of transcriptional changes in C. albicans biofilm-associated genes upon exposure to Sc3458. Gene expression shown as Log2 fold change, with significance (Log2 > 1 or <−1, P < 0.05; Statistical analysis was assessed using two-sided Wald tests with Benjamini–Hochberg FDR correction. Significant differences are indicated as *P < 0.05). B Representative images of C. albicans biofilms CA3153 on VECs treated with Sc3458 and the industrial control (IC, CNCM I-3856), shown in head and side views. Scale bars = 100 µm. C Relative C. albicans biofilm biomass at early (24 hpi) and mature (48 hpi) stages (n = 3) and D C. albicans Biofilm thickness on VECs at 24 hpi comparing conditions with and without Sc3458 or the industrial control (IC, CNCM I-3856; n = 4). Biomass is normalized to C. albicans alone (100%). E Redistribution of C. albicans (SC5314, CA3153, and I49) between biofilm and supernatant fractions during treatment with Sc3458 or with the industrial control (IC, CNCM I-3856; n = 3). F Redistribution of Sc3458 when exposed to C. albicans (SC5314, CA3153, and I49) biofilm (n = 3). G Relative C. albicans biofilm biomass with Sc3458 or with the industrial control (IC, CNCM I-3856) in direct contact or physically separated using transwell inserts (0.4 µm pore size) or exposed to heat-inactivated S. cerevisiae (n = 3). H Schematic model illustrating the proposed mechanisms by which S. cerevisiae (purple cells) inhibits C. albicans biofilm formation (beige cells) on VECs (pink layer). The inhibitory effects include nutrient sequestration, metabolic adaptation by S. cerevisiae and physical constraints that disrupt C. albicans adhesion, filamentation, and biofilm structural integrity. C–G Bars represent mean ± standard deviation (SD) from independent experiments, with dots representing the mean of technical replicates in the individual experiments. Statistical significance was determined using two-sided one-sample t-tests set at 100% (C) or one-way ANOVA with Holm-Šidák post hoc tests to compare treatment conditions within each C. albicans strain (D–G). Significant differences are depicted as the P-value. Schematic overview in panel (H) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. S. cerevisiae reduces C. albicans biofilm biomass, alters gene expression, and disrupts biofilm integrity. A Heatmap of transcriptional changes in C. albicans biofilm-associated genes upon exposure to Sc3458. Gene expression shown as Log2 fold change, with significance (Log2 > 1 or <−1, P < 0.05; Statistical analysis was assessed using two-sided Wald tests with Benjamini–Hochberg FDR correction. Significant differences are indicated as *P < 0.05). B Representative images of C. albicans biofilms CA3153 on VECs treated with Sc3458 and the industrial control (IC, CNCM I-3856), shown in head and side views. Scale bars = 100 µm. C Relative C. albicans biofilm biomass at early (24 hpi) and mature (48 hpi) stages (n = 3) and D C. albicans Biofilm thickness on VECs at 24 hpi comparing conditions with and without Sc3458 or the industrial control (IC, CNCM I-3856; n = 4). Biomass is normalized to C. albicans alone (100%). E Redistribution of C. albicans (SC5314, CA3153, and I49) between biofilm and supernatant fractions during treatment with Sc3458 or with the industrial control (IC, CNCM I-3856; n = 3). F Redistribution of Sc3458 when exposed to C. albicans (SC5314, CA3153, and I49) biofilm (n = 3). G Relative C. albicans biofilm biomass with Sc3458 or with the industrial control (IC, CNCM I-3856) in direct contact or physically separated using transwell inserts (0.4 µm pore size) or exposed to heat-inactivated S. cerevisiae (n = 3). H Schematic model illustrating the proposed mechanisms by which S. cerevisiae (purple cells) inhibits C. albicans biofilm formation (beige cells) on VECs (pink layer). The inhibitory effects include nutrient sequestration, metabolic adaptation by S. cerevisiae and physical constraints that disrupt C. albicans adhesion, filamentation, and biofilm structural integrity. C–G Bars represent mean ± standard deviation (SD) from independent experiments, with dots representing the mean of technical replicates in the individual experiments. Statistical significance was determined using two-sided one-sample t-tests set at 100% (C) or one-way ANOVA with Holm-Šidák post hoc tests to compare treatment conditions within each C. albicans strain (D–G). Significant differences are depicted as the P-value. Schematic overview in panel (H) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. Given the gene expression patterns in C. albicans indicating cell aggregation and metabolic stress, we investigated whether the inhibitory effects of Sc3458 depend on direct contact and metabolically active cells. Using transwells to separate Sc3458 from C. albicans, allowing nutrient and molecule exchange, completely abolished the inhibition of biofilm formation (Fig. 4G). Similarly, heat-inactivated Sc3458 cells failed to inhibit biofilm formation (Fig. 4G). These findings indicate that biofilm inhibition by Sc3458 depends on both direct contact and cellular metabolism. The IC strain showed similar contact-dependent behavior (Fig. 4G), reinforcing the robustness of this mechanism (Fig.4H). S. cerevisiae suppresses pro-inflammatory responses and neutrophil activation To assess how Sc3458 treatment modulates inflammatory responses, we analyzed epithelial transcriptional changes, revealing significant reprogramming at 24 hpi (Fig. 3B, C). A total of 1581 genes were downregulated and 267 upregulated (Fig. 5A). GO enrichment analysis identified upregulated biological processes, including mitochondrial function, DNA repair, and RNA transport and processing, in response to Sc3458 treatment (Fig. 5C). Although direct and well-established links to VVC pathology remain elusive, the GO-terms suggest a general activation of integral cellular maintenance and adaptive response pathways.Fig. 5S. cerevisiae strain 3458 modulates epithelial transcriptional responses and reduces neutrophil activation induced by C. albicans.A Volcano plot showing significantly upregulated (red) and downregulated (blue) genes in VECs infected with C. albicans at 24 hpi in the presence of Sc3458. B, C Gene Ontology (GO) enrichment analysis of downregulated (B, blue) and upregulated (C, red) genes in VECs, based on Log2 fold change > 1 or <−1 and P < 0.1. Fold enrichment is visualized as a color gradient. Data are from n = 3 biological replicates. GO enrichment analysis (B, C) was conducted using a two-sided hypergeometric test with Bonferroni correction for multiple testing (adjusted P < 0.05). D IL-1β release by hMDMs (n = 9 donors) infected with C. albicans-with or without S. cerevisiae. Horizontal lines represent values for the Sc3458-only (blue) condition, the industrial control-only (light blue; IC, CNCM I-3856) negative control (gray), included as a reference. E Schematic showing neutrophil stimulation (green) with supernatants from C. albicans-infected VECs (pink), treatment with S. cerevisiae (purple), collected after 24 h. Beige cells represent C. albicans, illustrating the indirect effect of the treatment on neutrophil responses. F Neutrophil IL-8 release (n = 3 donors) and G, H neutrophil ROS release (n = 5 donors). Horizontal lines represent values for the Sc3458-only (blue) condition, the industrial control-only (light blue; IC, CNCM I-3856) negative control (gray), included as a reference. I Heatmap of neutrophil activation markers (CD11b, CD66b, CD62L, CXCR2). Mean fluorescence intensity (MFI) is normalized to unstimulated controls. Data are from n = 5 donors. J Neutrophil chemotaxis. Data from n = 4 donors. Data are shown as individual donor values with bars representing mean ± SEM. Statistical significance was determined using two-way ANOVA with Holm-Šidák post hoc tests to compare treatment conditions within each C. albicans strain (D–H, J) and one-way ANOVA with Holm-Šidák post hoc tests to compare the C. albicans only (Ca only) condition (I). Significant differences are indicated as the P-value or *P < 0.05, **P < 0.01, ***P < 0.001 (I). Schematic overview in panel (E) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. S. cerevisiae strain 3458 modulates epithelial transcriptional responses and reduces neutrophil activation induced by C. albicans. A Volcano plot showing significantly upregulated (red) and downregulated (blue) genes in VECs infected with C. albicans at 24 hpi in the presence of Sc3458. B, C Gene Ontology (GO) enrichment analysis of downregulated (B, blue) and upregulated (C, red) genes in VECs, based on Log2 fold change > 1 or <−1 and P < 0.1. Fold enrichment is visualized as a color gradient. Data are from n = 3 biological replicates. GO enrichment analysis (B, C) was conducted using a two-sided hypergeometric test with Bonferroni correction for multiple testing (adjusted P < 0.05). D IL-1β release by hMDMs (n = 9 donors) infected with C. albicans-with or without S. cerevisiae. Horizontal lines represent values for the Sc3458-only (blue) condition, the industrial control-only (light blue; IC, CNCM I-3856) negative control (gray), included as a reference. E Schematic showing neutrophil stimulation (green) with supernatants from C. albicans-infected VECs (pink), treatment with S. cerevisiae (purple), collected after 24 h. Beige cells represent C. albicans, illustrating the indirect effect of the treatment on neutrophil responses. F Neutrophil IL-8 release (n = 3 donors) and G, H neutrophil ROS release (n = 5 donors). Horizontal lines represent values for the Sc3458-only (blue) condition, the industrial control-only (light blue; IC, CNCM I-3856) negative control (gray), included as a reference. I Heatmap of neutrophil activation markers (CD11b, CD66b, CD62L, CXCR2). Mean fluorescence intensity (MFI) is normalized to unstimulated controls. Data are from n = 5 donors. J Neutrophil chemotaxis. Data from n = 4 donors. Data are shown as individual donor values with bars representing mean ± SEM. Statistical significance was determined using two-way ANOVA with Holm-Šidák post hoc tests to compare treatment conditions within each C. albicans strain (D–H, J) and one-way ANOVA with Holm-Šidák post hoc tests to compare the C. albicans only (Ca only) condition (I). Significant differences are indicated as the P-value or *P < 0.05, **P < 0.01, ***P < 0.001 (I). Schematic overview in panel (E) was “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. The GO enrichment analysis highlighted many downregulated processes (Fig. S5). We focused particularly on those relevant to VVC, such as responses to stimuli, cell death, and inflammatory responses (Fig. 5B). Key affected pathways included “cell surface receptor signaling”, “environmental and abiotic stimulus responses”, and “regulation of apoptotic and inflammatory signaling”. Notably, critical immune pathways—Nuclear Factor kappa B (NF-κB), Mitogen-Activated Protein Kinase (MAPK), and Phosphoinositide 3-Kinase/Protein Kinase B (PI3K/Akt) cascades—were downregulated, alongside Wnt signaling, phagocytosis, and autophagy regulation. Given the importance of MAPK, NF-κB, and PI3K/Akt pathways in epithelial responses to C. albicans29,33, we further analyzed specific DEGs within these pathways. Key genes in the MAPK pathway (MKK7 and MLK3) were downregulated, critical for activating JNK and p38 (Fig. S6)34. MK2 and ICAM1, involved in inflammatory signaling, were also downregulated. MK2 regulates cytokine production via the p38 MAPK pathway, while ICAM1 facilitates immune cell adhesion and recruitment35. ERK1, part of a separate MAPK branch, was also downregulated, a pathway associated with processes such as cell proliferation and stress responses to C. albicans36. In the NF-κB pathway, p65, RelB, and NEMO were all downregulated, components essential for cytokine transcription and anti-apoptotic signaling37. Similarly, AKT1, AKT2, and IRS1, part of the PI3K/Akt pathway, were downregulated, affecting survival signaling38. Reduced expression of TLR2 and FGFR1/4 indicated diminished immune activation and growth factor signaling. Downregulation of eNOS, FOXO1, and FOXO4 further supports a shift in apoptosis regulation39. Transcriptional profiling revealed upregulation of IL1B in C. albicans-infected VECs exposed to S. cerevisiae (Fig. S6). Excessive IL-1β production hyperactivates neutrophils, causing tissue damage without effectively clearing the pathogen, a hallmark of severe VVC. Unlike macrophages, VECs produce limited inflammasome-dependent IL-1β, and no IL-1β responses have been detected upon C. albicans infection of VECs40. By assessing IL-1β release by human monocyte-derived macrophages (hMDMs)8, we found that Sc3458 tended to reduce IL-1β responses to C. albicans (Fig. 5D). In contrast, the IC strain increased IL-1β secretion (Fig. 5D), indicating that it reinforces inflammation across both epithelial (Fig. 2) and macrophage compartments. These data highlight a key functional distinction: while IC broadly amplifies inflammation, Sc3458 appears to reduce inflammation across epithelial cells and macrophages. Next, we validated how the downregulation of pro-inflammatory MAPK and NF-κB pathways corresponded to reduced downstream neutrophil activation by the epithelial cells. Supernatants from epithelial infections with Sc3458 led to slightly lower IL-8 secretion (Fig. 5E, F) and reduced ROS release by neutrophils (Fig. 5G). Both effects were also seen with the IC strain, albeit with strain-specific variability. Although not all changes were statistically significant, an overall trend of reduced neutrophil activation was evident. In line with this, neutrophils exposed to supernatants from C. albicans-infected VECs treated with Sc3458 exhibited altered surface marker expression (Fig. 5H). CD11b, an integrin involved in fungal recognition and recruitment, was downregulated, alongside CD66b, a marker of secondary granule mobilization. CD35 also decreased, suggesting reduced degranulation. In contrast, CD16 expression was elevated, consistent with enhanced neutrophil viability or antimicrobial readiness. Notably, CD62L, typically shed during activation, was preserved, indicating a less activated phenotype. Moderate reduction in CXCR2, which is internalized upon chemokine engagement, pointed to recent activation and potential migratory readiness. These expression patterns were broadly comparable to those induced by the IC strain, suggesting that, despite differences in upstream modulation of inflammation, both strains elicit a similarly tempered neutrophil phenotype. Nevertheless, responses varied depending on the C. albicans isolate used, with CA3153 and I49 generally inducing weaker pro-inflammatory signaling. Importantly, Sc3458 and IC did not affect neutrophil chemotaxis across the endothelial-epithelial barrier in response to C. albicans-infected VECs (Fig. 5I). Collectively, these findings indicate that Sc3458 modulates neutrophil activation by dampening upstream epithelial signaling, while preserving neutrophil recruitment to the site of infection. Sc3458 reduces epithelial inflammation primarily via secondary modulation of pathogenic stimuli To disentangle direct epithelial effects of Sc3458 from secondary consequences of altered C. albicans pathogenicity, we profiled transcriptional changes in VECs when treated with Sc3458 in the absence of C. albicans infection. This is to dissect the epithelial responses to Sc3458 independent of C. albicans burden, hyphal formation, or epithelial damage (Fig. S7A). PCA revealed distinct clustering between Sc3458-treated and untreated VECs at both 6 and 24 hpi, indicating time-dependent transcriptional reprogramming (Fig. S7B). At 6 hpi, global transcriptomic profiles diverged without individual genes reaching differential expression thresholds, suggesting subtle but coordinated early epithelial priming (Fig. S7B, C). By 24 hpi, Sc3458 induced a defined transcriptional response, with 181 differentially expressed genes (153 upregulated, 28 downregulated; Fig. S7B–D). Gene Ontology enrichment revealed downregulation of hormone-responsive, lipid metabolic, and proliferative pathways, indicative of a hormonally quiescent, less proliferative epithelial state (Fig. S7E). In contrast, pathways associated with the unfolded protein response and endoplasmic reticulum stress adaptation were significantly enriched, reflecting activation of tightly regulated, cytoprotective stress responses rather than cytotoxicity (Fig. S7F). These transcriptional changes suggest that Sc3458 promotes a cytoprotective epithelial state marked by hormonal quiescence, reduced proliferation, and enhanced proteostasis. Notably, despite previously observed attenuation of NF-κB and MAPK signaling in C. albicans-infected epithelial cells (Fig. 5B, Fig. S6), no differential expression of genes associated with these pathways was detected in VECs exposed to Sc3458 alone. Together, these data support that Sc3458 reduces epithelial inflammation via strain-specific epithelial stabilization combined with infection-dependent attenuation of pathogenic stimuli, reinforcing epithelial resilience rather than directly targeting classical inflammatory signaling pathways. S. cerevisiae promotes neutrophil-mediated C. albicans clearance Given that Sc3458 reduced neutrophil activation downstream of C. albicans-infected VECs, we next assessed its direct effects on neutrophil-mediated fungal clearance, known to be impaired during VVC8. In addition to inhibiting C. albicans biofilm formation (Fig. 4), Sc3458 treatment enhanced neutrophil-mediated clearance of C. albicans biofilm-associated cells (Fig. 6A). This led to an additive effect, with C. albicans biofilm biomass reduced by 30–50% beyond the inhibition observed with Sc3458 treatment alone. To further evaluate this interaction, we assessed neutrophil-mediated killing at early time points. Sc3458 significantly enhanced fungal killing at both 1 and 3 hpi, reducing fungal viability up to ~40% compared to neutrophils alone (Fig. 6B). A similar enhancement of neutrophil-mediated clearance was observed with strain IC, suggesting that this trait can be shared across S. cerevisiae strains. The enhanced killing was consistent for C. albicans isolates SC5314 and CA3153. However, for isolate I49, which previously showed reduced virulence and delayed immune activation, the neutrophil-mediated killing effect was absent at 1 hpi. To determine whether S. cerevisiae itself constituted a target of neutrophil-mediated clearance under these conditions, we directly quantified neutrophil-mediated killing of Sc3458 and IC. In contrast to C. albicans, neither S. cerevisiae strain was efficiently killed by human neutrophils over 3 h, with CFU counts remaining stable or even exceeding the inoculum (Fig. S9). Notably, co-culture with C. albicans did not affect S. cerevisiae survival, indicating that neutrophils preferentially show an increased clearance of C. albicans without compromising Sc3458 viability in mixed fungal settings.Fig. 6Neutrophil-mediated killing, biofilm inhibition, and NETosis of C. albicans in the presence of S. cerevisiae.A Neutrophil-mediated killing of C. albicans biofilms (strains SC5314 (+ Sc34458 n = 10, + IC n = 9), CA3153 (+ Sc34458 n = 7, + IC n = 9), and I49 (+ Sc34458 n = 9, + IC n = 9)) in the presence of Sc3458 or the industrial control (IC, CNCM I-3856) at 3 hpi. Data are normalized to the inhibitory effects of S. cerevisiae alone and compared to neutrophils with C. albicans alone. B Neutrophil-mediated killing of C. albicans (strains SC5314, CA3153, and I49) assessed at 1 (n = 10) and 3 hpi (SC5314 (+ Sc34458 n = 9, + IC n = 9), CA3153 (+ Sc34458 n = 8, + IC n = 10), and I49 (+ Sc34458 n = 9, + IC n = 10))with Sc3458 or with the industrial control (IC, CNCM I-3856). Data are normalized to C. albicans with neutrophils alone (100%). C Representative images of C. albicans (strain CA3153) treated with neutrophils in the presence of Sc3458. Images show C. albicans (red), neutrophil death/NETosis via Sytox Green staining (green), and merged phase contrast. Scale bars = 200 μm. D Overgrowth of C. albicans strains SC5314, CA3153, and I49 in the presence of neutrophils, with or without S. cerevisiae strain 3458 or the industrial control (IC, CNCM I-3856), and quantified by Total Orange Object Integrated Intensity. E NETosis induced by C. albicans strains SC5314 (left), CA3153 (middle), and I49 (right) under the same conditions as in (D). NETosis was quantified via Sytox Green fluorescence and levels were normalized to the total neutrophil count. Data are shown as individual donor values with bars representing mean ± SEM and dots representing individual donors. Statistical significance was determined using two-sided one-sample t-tests against a mean of 100 (A, B). Statistical significance was determined using a linear mixed effects model with a Holm-Šidák multiple comparisons test (D, E). Significance is depicted as the P-value or * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (D, E). Neutrophil-mediated killing, biofilm inhibition, and NETosis of C. albicans in the presence of S. cerevisiae. A Neutrophil-mediated killing of C. albicans biofilms (strains SC5314 (+ Sc34458 n = 10, + IC n = 9), CA3153 (+ Sc34458 n = 7, + IC n = 9), and I49 (+ Sc34458 n = 9, + IC n = 9)) in the presence of Sc3458 or the industrial control (IC, CNCM I-3856) at 3 hpi. Data are normalized to the inhibitory effects of S. cerevisiae alone and compared to neutrophils with C. albicans alone. B Neutrophil-mediated killing of C. albicans (strains SC5314, CA3153, and I49) assessed at 1 (n = 10) and 3 hpi (SC5314 (+ Sc34458 n = 9, + IC n = 9), CA3153 (+ Sc34458 n = 8, + IC n = 10), and I49 (+ Sc34458 n = 9, + IC n = 10))with Sc3458 or with the industrial control (IC, CNCM I-3856). Data are normalized to C. albicans with neutrophils alone (100%). C Representative images of C. albicans (strain CA3153) treated with neutrophils in the presence of Sc3458. Images show C. albicans (red), neutrophil death/NETosis via Sytox Green staining (green), and merged phase contrast. Scale bars = 200 μm. D Overgrowth of C. albicans strains SC5314, CA3153, and I49 in the presence of neutrophils, with or without S. cerevisiae strain 3458 or the industrial control (IC, CNCM I-3856), and quantified by Total Orange Object Integrated Intensity. E NETosis induced by C. albicans strains SC5314 (left), CA3153 (middle), and I49 (right) under the same conditions as in (D). NETosis was quantified via Sytox Green fluorescence and levels were normalized to the total neutrophil count. Data are shown as individual donor values with bars representing mean ± SEM and dots representing individual donors. Statistical significance was determined using two-sided one-sample t-tests against a mean of 100 (A, B). Statistical significance was determined using a linear mixed effects model with a Holm-Šidák multiple comparisons test (D, E). Significance is depicted as the P-value or * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 (D, E). Using live-cell analysis, we simultaneously assessed fungal overgrowth, neutrophil death and NETosis (Fig. 6C). Quantification of mScarlet-tagged C. albicans revealed significant reductions in C. albicans burden upon Sc3458 and IC treatment overtime, starting 2 hpi (Fig. 6D). The visualization of treatment after 6 h helped us to understand the interactions between neutrophils and C. albicans, in the presence of S. cerevisiae (Fig. 6C). Images show enhanced neutrophil clustering and reduced C. albicans overgrowth upon Sc3458 treatment compared to neutrophils cultured with C. albicans alone, where fungal persistence was more evident. Interestingly, improved C. albicans killing and reduced overgrowth were not accompanied by significant changes in overall neutrophil cell death (Fig. S8). Similarly, NETosis events were not enhanced by the presence of Sc3458 or strain IC (Fig. 6E), suggesting that enhanced fungal clearance occurs through mechanisms independent of NETosis. S. cerevisiae 3458 reduces C. albicans burden and attenuates local inflammation in a murine model To determine whether the pathogenicity attenuating, and immunomodulatory effects observed in vitro translated to an improved disease outcome, we treated mice with Sc3458 in the context of a well-established murine VVC model. Mice in a pseudoestrous state were intravaginally administered Sc3458, the industrial control strain (IC), or PBS (10 µL of 108 cells/mL per mouse) one day after infection with C. albicans CA3153 or SC5314 (10 µL of 107 cells/mL per mouse) (Fig. 7A). To quantify vaginal C. albicans burden, longitudinal bioluminescence imaging (BLI) was used, which minimizes disruption of the infection in contrast to repeated vaginal lavages for CFU assessment (Fig. S10). BLI quantification revealed a significant reduction in C. albicans burden following treatment with Sc3458. From day 6 onward, mice treated with Sc3458 exhibited a sustained decrease in C. albicans for both CA3153 (Fig. 7B, D; Fig. S11) and SC5314 (Fig. 7C, E; Fig. S11) vaginal burden, while PBS-treated controls increased in fungal burden. The IC strain showed a slight early reduction in SC5314 fungal burden compared to C. albicans alone; however, this trend was not statistically significant, highlighting a lack of durable anti-Candida activity compared to Sc3458. CFU enumeration at day 10 supported the trends observed by BLI, with Sc3458-treated animals generally exhibiting lower fungal burdens than control groups; however, variability within groups precluded statistical significance (Fig. S12).Fig. 7Efficacy of S. cerevisiae Sc3458 in diminishing C. albicans load in an in vivo murine model of VVC.A Schematic representation of the experimental setup, including estradiol pre-treatment, vaginal infection with C. albicans (CA3153 or SC5314), and subsequent intravaginal administration of Sc3458, the industrial control strain (IC; CNCM I-3856), or PBS. B, C In vivo bioluminescence imaging (BLI)-based quantification of fungal burden, expressed as total photon flux, in C. albicans-infected mice (B: CA3153; C: SC5314) treated with Sc3458 (B n = 12, C n = 8), IC (B n = 11, C n = 8), or PBS (B n = 11, C n = 8). BLI measurements were acquired on days 2, 4, 6, 8, and 10 post-infection. Photon flux values were calculated using a region of interest (ROI) covering the entire vaginal area. D, E Representative BLI images of mice from each treatment group at day 10 post-infection, corresponding to panel (B) for CA3153 (D) and panel (C) for SC5314 (E). Data are presented as mean ± SEM, with individual data points representing single animals. Statistical comparisons were performed using one-way ANOVA followed by Holm-Šidák multiple-comparisons testing, with comparisons made against the C. albicans-infected PBS control group. Efficacy of S. cerevisiae Sc3458 in diminishing C. albicans load in an in vivo murine model of VVC. A Schematic representation of the experimental setup, including estradiol pre-treatment, vaginal infection with C. albicans (CA3153 or SC5314), and subsequent intravaginal administration of Sc3458, the industrial control strain (IC; CNCM I-3856), or PBS. B, C In vivo bioluminescence imaging (BLI)-based quantification of fungal burden, expressed as total photon flux, in C. albicans-infected mice (B: CA3153; C: SC5314) treated with Sc3458 (B n = 12, C n = 8), IC (B n = 11, C n = 8), or PBS (B n = 11, C n = 8). BLI measurements were acquired on days 2, 4, 6, 8, and 10 post-infection. Photon flux values were calculated using a region of interest (ROI) covering the entire vaginal area. D, E Representative BLI images of mice from each treatment group at day 10 post-infection, corresponding to panel (B) for CA3153 (D) and panel (C) for SC5314 (E). Data are presented as mean ± SEM, with individual data points representing single animals. Statistical comparisons were performed using one-way ANOVA followed by Holm-Šidák multiple-comparisons testing, with comparisons made against the C. albicans-infected PBS control group. Along with the high fungal burden of CA3153 observed in PBS-treated controls 10 days post-infection, we detected pronounced IL-1-responses in vaginal lavages, characterized by elevated levels of both IL-1β and a trend towards increased IL-1α. IL-1-mediated hyperinflammation has been extensively associated with the dysregulated inflammatory state that underlies disease severity in VVC41. Sc3458-treated mice exhibited significantly reduced IL-1β levels (Fig. 8A). Consistent with this observation chemokines CXCL2, CCL3 (MIP-1α), and CCL4 (MIP-1β) involved in innate immune cell recruitment, were significantly reduced in the Sc3458-treated mice. Given the 10-day infection, adaptive immune responses are also activated. Although controversial regarding their contribution to VVC8, treatment with Sc3458 significantly reduced both Th17- and Th22-associated cytokines IL-17A and IL-22 (Fig. 8A). Furthermore, type 2 immunity-related cytokines, including IL-4 and the eosinophil-recruiting chemokine CCL11 (eotaxin-1), were not significantly induced in the Sc3458-treated group (Fig. 8A, Fig. S13), suggesting potentially a lower level of mucosal irritation and symptom severity. This indicates an overall tendency towards dampened inflammation in the vaginal mucosa of mice treated with Sc3458 (Fig. 8A, Fig. S13). While the industrial control also decreased levels of several inflammatory mediators (CXCL2, CCL3, and CCL4), it failed to dampen IL-1β release as well as type 17 (IL-17 and IL-22) and type 2 (IL-4 and CCL11) responses (Fig. 8A, Fig. S13). These observations reaffirm the unique immunomodulatory potential of Sc3458 compared to the IC strain.Fig. 8Efficacy of S. cerevisiae Sc3458 in diminishing C. albicans CA3153-induced key inflammatory mediators in an in vivo murine model of VVC.A Cytokine and chemokine analysis of vaginal lavage fluids collected on day 10 from C. albicans CA3153-infected mice (n = 11) treated with Sc3458 (n = 12), industrial control strain (IC, CNCM I-3856; n = 11), or PBS (n = 14). Data include cytokines involved in: initiation of inflammation (IL-1α, IL-1β), innate immune recruitment (CXCL2, CCL3 MIP-1α, CCL4, and MIP-1β), T-cell associated cytokines (IL-17A, IL-22), and allergy-related markers (IL-4, CCL11). PBS-treated mice serve as uninfected baseline controls. B Absolute numbers of neutrophils, macrophages, and eosinophils in vaginal lavage fluids from each treatment group on day 10 post-infection. CA3153-infected mice (n = 11) treated with Sc3458 (n = 11), industrial control strain (IC, CNCM I-3856; n = 10), or PBS (n = 11). PBS-treated mice served as uninfected baseline controls. C Heatmap representation of normalized expression levels (median fluorescence intensity) of activation markers CD11b, CD11c, Ly6C, and Ly6G on eosinophils, compared to the PBS-only condition. D Representative histological assessment of the vaginal epithelium using hematoxylin and eosin (H&E) staining in pseudoestrus-induced mice infected with C. albicans CA3153 and treated with Sc3458, the industrial control strain (IC, CNCM I-3856), or PBS from two independent experiments. E Histological inflammation scores in vaginal tissue following treatment, assessed by a blinded experimenter using a standardized scoring system (0–4) based on neutrophil infiltration and tissue damage. Scores represent the average of at least 10 fields across proximal, mid, and distal regions imaged at 20× magnification (n = 8). (A, B, E) Bars represent mean ± SEM. Statistical analysis: Depending on data distribution, statistical comparisons were performed using either one-way ANOVA followed by Holm-Šidák multiple comparisons test (A, B, E) or Kruskal-Wallis test followed by Dunn’s multiple comparisons test (C). Comparisons were made against the C. albicans-infected, PBS-treated control group (A, B) and against the uninfected, PBS-treated group (C). Significance is depicted as exact P-values (A, E) or * P < 0.05 (C). Efficacy of S. cerevisiae Sc3458 in diminishing C. albicans CA3153-induced key inflammatory mediators in an in vivo murine model of VVC. A Cytokine and chemokine analysis of vaginal lavage fluids collected on day 10 from C. albicans CA3153-infected mice (n = 11) treated with Sc3458 (n = 12), industrial control strain (IC, CNCM I-3856; n = 11), or PBS (n = 14). Data include cytokines involved in: initiation of inflammation (IL-1α, IL-1β), innate immune recruitment (CXCL2, CCL3 MIP-1α, CCL4, and MIP-1β), T-cell associated cytokines (IL-17A, IL-22), and allergy-related markers (IL-4, CCL11). PBS-treated mice serve as uninfected baseline controls. B Absolute numbers of neutrophils, macrophages, and eosinophils in vaginal lavage fluids from each treatment group on day 10 post-infection. CA3153-infected mice (n = 11) treated with Sc3458 (n = 11), industrial control strain (IC, CNCM I-3856; n = 10), or PBS (n = 11). PBS-treated mice served as uninfected baseline controls. C Heatmap representation of normalized expression levels (median fluorescence intensity) of activation markers CD11b, CD11c, Ly6C, and Ly6G on eosinophils, compared to the PBS-only condition. D Representative histological assessment of the vaginal epithelium using hematoxylin and eosin (H&E) staining in pseudoestrus-induced mice infected with C. albicans CA3153 and treated with Sc3458, the industrial control strain (IC, CNCM I-3856), or PBS from two independent experiments. E Histological inflammation scores in vaginal tissue following treatment, assessed by a blinded experimenter using a standardized scoring system (0–4) based on neutrophil infiltration and tissue damage. Scores represent the average of at least 10 fields across proximal, mid, and distal regions imaged at 20× magnification (n = 8). (A, B, E) Bars represent mean ± SEM. Statistical analysis: Depending on data distribution, statistical comparisons were performed using either one-way ANOVA followed by Holm-Šidák multiple comparisons test (A, B, E) or Kruskal-Wallis test followed by Dunn’s multiple comparisons test (C). Comparisons were made against the C. albicans-infected, PBS-treated control group (A, B) and against the uninfected, PBS-treated group (C). Significance is depicted as exact P-values (A, E) or * P < 0.05 (C). In contrast to the pronounced immunomodulatory effects observed during CA3153 infection, Sc3458 treatment did not substantially alter the inflammatory landscape during infection with C. albicans SC5314. Although Sc3458 treatment resulted in a clear reduction in vaginal fungal burden at 10 days post-infection, this was not accompanied by a marked attenuation of IL-1-driven inflammation or downstream chemokine and cytokine responses (Figs. S14 and 15). Levels of IL-1α, IL-1β, CXCL2, CCL3, and CCL4, as well as Th17- and type 2-associated cytokines, remained largely comparable to those observed in SC5314-infected controls. Flow cytometry on vaginal lavages from CA3153 and SC5314 infected mice revealed minimal reductions in total neutrophil, macrophage and eosinophil counts upon Sc3458 treatment (Fig. 8B, Fig. S14B). However, as Sc3458-treated mice exhibited reduced fungal burden, this could indicate enhanced neutrophil functionality despite similar cell numbers. Although Sc3458 treatment only a modestly affected eosinophil abundance, reduced expression of the integrin CD11c (CA3153 and SC5314) and CD11b (SC5314) suggest altered eosinophil activation and tissue interaction during C. albicans infection (Fig. 8C, Fig. S14B). While these integrins are often used as activation markers, they also play direct roles in pathogen recognition and cytokine signaling in eosinophils42. The observed downregulation is consistent with reduced type 2 inflammatory cytokines and may reflect a shift toward a less activated, less tissue-damaging phenotype. To obtain insights beyond neutrophil presence in vaginal lavages, histological analysis was performed to examine the state of the vaginal epithelium. Mice infected with CA3153 or SC5314 displayed pronounced neutrophil infiltration in the superficial epithelial layers lining the vaginal lumen, together with large aggregates of neutrophils and shed epithelial cells within the lumen (Fig. 8D). In localized regions, abscess-like structures containing large numbers of neutrophils were observed, some of which appeared ruptured, releasing neutrophils and epithelial debris into the vaginal lumen. In CA3153-infected mice treated with Sc3458, the inflammation was scored significantly lower (Fig. 8E), based on the observation that the vaginal epithelium was largely devoid of infiltrating neutrophils and microabscesses, although some neutrophils could be observed in the vaginal lumen. In contrast, IC-treated mice continued to display neutrophils embedded within the epithelial barrier albeit at lower numbers then CA3153 treated mice. Similarly, in SC5314-infected mice, Sc3458 treatment was associated with a significantly reduced neutrophil infiltration score (Fig. 8E), although this reduction was less pronounced than that observed in CA3153-infected animals. By comparison, IC-treated mice still exhibited persistent neutrophil accumulation within the epithelial layer. Discussion This study establishes the therapeutic potential of a specific S. cerevisiae strain in the context of C. albicans-driven VVC. Through systematic screening, we identified S. cerevisiae 3458, originally isolated from the vaginal mycobiota, as highly capable of inhibiting C. albicans virulence traits. This strain further set itself apart in the ability to reduce epithelial damage and inflammatory responses. Transcriptomic profiling and in vitro infection models revealed that Sc3458 attenuates C. albicans biofilm formation and suppresses pro-inflammatory epithelial signaling (MAPK, NF-κB). Additionally, it enhanced neutrophil-mediated fungal clearance in vitro. In a murine model of VVC, Sc3458 treatment significantly reduced fungal burden while alleviating IL-1β-driven hyperinflammation, reducing innate chemokines, as well as dampening type-17 and type-2 responses. Collectively, this underscores Sc3458’s translational relevance as a live biotherapeutic candidate capable of preventing disease development and promoting resolution of inflammation. Effective and affordable treatments for VVC are urgently needed due to its high prevalence and the increasing incidence of treatment failures driven by rising antifungal resistance6,43,44. Nevertheless, a universal cure is challenging due to heterogeneous predisposing factors, including genetic predisposition affecting immune responses, C. albicans virulence, and personal habits45. Despite this complexity, the prevailing view is that while hyperactivation of the epithelial cells and neutrophils play a major role in the onset of VVC, C. albicans virulence is equally crucial in driving disease progression10. To address these multifactorial drivers, we explored S. cerevisiae as a live biotherapeutic targeting multiple VVC-related factors23,46. S. cerevisiae has proven itself as a promising probiotic across diverse niches, combating pathogens and modulating immune responses23,25,46,47. However, our screening revealed that previously characterized probiotic isolates were not effective as live biotherapeutics against all tested C. albicans VVC-related factors. No clear correlation between niche specificity and antifungal activity emerged, likely due to strain-specific differences shaped by habitat-related idiosyncrasies, such as the microbial environment, host factors, and nutrient availability48,49. While the two most effective isolates, 3458 and 3471, originated from vaginal environments, other vaginal isolates showed limited efficacy. This underscores that habitat origin alone does not predict performance as a live biotherapeutic, reflecting the complexity of strain-specific adaptations. Sc3458 potently attenuated multiple C. albicans virulence traits, including proliferation, adhesion to epithelial surfaces, and the yeast-to-hypha transition. Given the multifactorial nature of effects against C. albicans and inflammation, it is unlikely that a single mechanism underlies this inhibition. Although early time points could capture direct responses, only a small number of genes were differentially expressed, and these are not established virulence factors. Genes such as CSA2, ADH2, and FMN1 are primarily involved in metabolic or housekeeping functions50,51. In contrast, the broader transcriptional changes at 24 hpi, offer more insight into cellular responses to S. cerevisiae treatment. Metabolic interference, previously described for lactobacilli, is a plausible contributor31,52. Transcriptomic profiling revealed upregulation of amino acid biosynthesis (MET1, MET2, LYS4, TRP4), alternative carbon metabolism (ACS1, GPD2), and the TCA cycle (ACO2, ACC1), consistent with nutrient stress in C. albicans. While S. cerevisiae is also known to secrete antifungal metabolites such as phenylethanol and tryptophol53,54, the inhibitory effects we observed were absent or markedly diminished when direct contact between organisms was prevented. Based on both transcriptomic and contact-dependence assays, we conclude that inhibition is not solely driven by diffusible metabolic products. Rather, contact appears to facilitate localized nutrient depletion. This contact dependence is a trait of S. cerevisiae previously linked to co-aggregation of S. cerevisiae and C. albicans through ALS gene-mediated interactions55, imposing physical constraints on C. albicans and intensifying metabolic competition. Additionally, the loss of inhibitory effects upon heat inactivation suggests that it depends on metabolically active cells. However, as heat inactivation also alters cell surface properties56, the observed C. albicans transcriptional reprogramming may result from both metabolic competition and altered cell-cell interactions, warranting further investigation. Building on this, S. cerevisiae 3458 interfered with C. albicans biofilm formation on epithelial cells through both physical disruption (downregulation of TRY5, SUC1, SAP8, HGC1, YAK1) and metabolic interference (downregulation of HGT13, GAL1, MAL31). S. cerevisiae integrated into C. albicans biofilms on epithelial cells, reducing biofilm density, and displaced C. albicans cells toward the vaginal lumen, where clearance is likely enhanced57,58. This resembles Lacticaseibacillus rhamnosus-mediated displacement of C. albicans from intestinal epithelia59. The downregulation of adhesion, morphogenesis, and biofilm genes in C. albicans suggests structural repression, likely driven by S. cerevisiae-induced metabolic competition. This supports a dual mechanism where metabolic and physical interference jointly disrupt biofilm integrity. By disrupting biofilm structural integrity, S. cerevisiae diminishes the biofilm’s protective properties against antifungals and immune cell infiltration60,61. This may enhance the efficacy of antifungal therapy, providing a potential avenue for synergistic treatment strategies62. While C. albicans virulence is a key factor in VVC pathology, dysregulated inflammatory responses drive symptom severity8,63. This is supported by elevated activation of key inflammatory signaling pathways (NF-κB, c-Fos, and p38) in both symptomatic and asymptomatic women colonized with hyphae-forming C. albicans strains8,64. We observed that reductions in C. albicans virulence by Sc3458 were accompanied by dampened epithelial IL-8 and GM-CSF responses. The reduced C. albicans colonization and hyphal morphology were associated with reduced activation of MAPK and NF-κB pathways, both critical for driving inflammatory responses by VECs29. Notably, S. cerevisiae has been shown to modulate these same pathways in other host environments25, suggesting a conserved mechanism of immune modulation. Within the NF-κB pathway, reduced expression of key components such as RELA (p65), RELB, and NEMO was observed, potentially impairing pro-inflammatory cytokine transcription29,65. Similarly, downregulation of MAPK components, including MKK7 and MLK3, suggests diminished JNK and p38 pathway activation66,67, while reduced ERK1 expression, associated with cell proliferation and stress responses, indicates decreased MAPK signaling overall. This reduction in MAPK signaling could increase the epithelial threshold to mount pro-inflammatory cytokine responses, potentially mitigating mucosal inflammation during VVC36,65. Furthermore, S. cerevisiae downregulated PI3K/Akt pathway components, including AKT1, AKT2, and IRS1, further supporting an overall dampening of epithelial inflammatory signaling. Downregulation of pattern recognition and growth factor receptors such as TLR2 and FGFR1/4, may further contribute to limiting immune activation and epithelial stress responses33. While transcriptional changes may not directly reflect protein activity, the transcriptional shifts to a less inflammatory state align well with reduced fungal burden, inhibition of the yeast-to-hypha transition, decreased biofilm formation, and dampened inflammatory cytokine release from VECs. This scenario likely lowers epithelial immune responses that downstream drive neutrophil activation in VVC8. Correspondingly, we observed reduced epithelial damage alongside decreased IL-8 and GM-CSF release. Similarly, monocyte-derived macrophages also showed a tendency towards reduced C. albicans-induced IL-1 responses in the presence of Sc3458. Given that these responses are critically dependent on the C. albicans toxin candidalysin40,68–70, it is plausible that S. cerevisiae interferes with toxin delivery, thereby mitigating cell death and associated inflammation27. While release of the cytoplasmic enzyme LDH can be used as a reliable readout for C. albicans-induced cell death in vitro, in vivo translation of increased LDH in the vaginal lumen has limitations. Particularly, since an increased epithelial turnover can lead to increased deposition of shed epithelial cells in the vaginal lumen that can release LDH. In addition to modulating epithelial responses during C. albicans infection, Sc3458 induced transcriptional changes in VECs in the absence of infection. Sc3458 promoted a homeostatic, non-inflammatory state characterized by downregulation of hormone-responsive and proliferation-associated pathways. As hormonal fluctuations and elevated epithelial turnover can compromise barrier integrity and exacerbate mucosal inflammation, this response may contribute to epithelial stabilization and enhanced barrier function71,72. Moreover, Sc3458 activated UPR and ER-stress pathways, indicative of an adaptive cytoprotective response. Consistent with this, prior induction of UPR suppressed epithelial cytotoxicity during C. glabrata infection under pathogenic conditions73, supporting a protective role of UPR in epithelial resistance to Candida species. Despite these transcriptional changes, Sc3458 alone did not induce strong modulation of NF-κB or MAPK signaling, in line with its modest direct immunostimulatory effects. Collectively, these findings suggest that Sc3458 enhances epithelial stability and stress resilience, while its immunomodulatory activity becomes most evident during C. albicans infection. Importantly, in vivo Sc3458 treatment during infection with the vaginal isolate CA3153 significantly suppressed IL-1β responses as well as inflammatory mediators involved in innate immune cell recruitment, including CXCL2, CCL3 (MIP-1α), and CCL4 (MIP-1β). This coordinated reduction in IL-1β and chemokine levels suggests attenuation of the feed-forward inflammatory cascade that promotes excessive leukocyte recruitment and activation that drives mucosal inflammation within the vaginal niche8. VVC is suggested to involve type 2 immunity, with IL-4 implicated in the mucosal allergy-like response seen in severe cases74. C. albicans CA3153 infection failed to significantly induce IL-4 and CCL11 responses upon Sc3458 treatment. Further, C. albicans infection induced significantly lower type 17 (IL-17 and IL-22) responses upon Sc3458 treatment, suggesting a suppression of Th17-mediated inflammation8. Although IL-17 and IL-22 are not strictly required for fungal control in all VVC models75 their overactivation may drive mucosal inflammation. Notably, controlled elevation of IL-22 has been linked to protection through antimicrobial peptide induction76. Collectively, Sc3458 may fine-tune inflammatory responses, limiting immunopathology while maintaining effective fungal clearance. In contrast, the industrial control strain failed to reduce IL-1β, IL-17, IL-22, IL-4, and CCL11 responses to C. albicans CA3153 infection. A potential limitation is, however, that Sc3458 showed less efficacy against the highly virulent reference C. albicans strain SC5314. This underscores, that S. cerevisiae-based live-biotherapeutics may have a limited spectrum of efficacy against specific C. albicans isolates, which warrant further validation against collections of vaginal C. albicans isolates. Although C. albicans load and epithelial signaling are reduced by S. cerevisiae, treatment is unlikely to begin immediately at infection onset, allowing time for dysregulated non-self-limiting neutrophil-mediated inflammation8. Consequently, S. cerevisiae’s effect on VVC pathology may remain limited without directly influencing neutrophil activation and efficacy. We observed reduced epithelial-mediated neutrophil activation in the presence of S. cerevisiae in vitro, including lower IL-8 and ROS levels, which are associated with heightened inflammation and exacerbated VVC symptoms74,77. Neutrophil surface marker expression suggested an altered activation state. Downregulation of CD35 and CD66b suggested reduced degranulation and release of tissue-damaging enzymes in response to C. albicans-infected VECs78. In contrast, CD62L, typically shed during activation, was preserved, indicating a less activated neutrophil phenotype79. Conversely, the chemokine receptor CXCR2 which is internalized upon chemokine sensing, showed moderate downregulation, supporting recent or active neutrophil recruitment80. While CD16, linked to antimicrobial functionality, was upregulated81. Importantly, CD11b, an integrin involved in neutrophil adhesion, recruitment, and recognition of C. albicans antigens such as Als3, Pra1 and β-glucan82–84, was also decreased, pointing to diminished activation potential by direct fungal recognition. Together, these findings highlight that S. cerevisiae can alter the neutrophil activation state while preserving its ability to effectively counteract the infection. In vivo, Sc3458 treatment reduced overall leukocyte numbers in vaginal lavages, with Sc3458-treated mice displaying the lowest counts of all measured myeloid cell populations among infected groups. Although not all differences reached statistical significance, this pattern suggests a general dampening of inflammatory cell infiltration. Consistently, histological analysis showed reduced neutrophil infiltration within the vaginal epithelium following treatment. Among infected mice, Sc3458-treated animals also exhibited the lowest integrin expression on eosinophils, indicating reduced inflammatory activation. The discrepancy between lavage and histological measurements likely reflects the dynamic release of neutrophils from the infected epithelium into the vaginal lumen. Together, these findings suggest that Sc3458 modulates, rather than suppresses, the inflammatory response, reducing pathological inflammation while maintaining granulocytes required for fungal control. Accordingly, S. cerevisiae significantly enhanced neutrophil killing of C. albicans in vitro, without inducing excessive NET formation or compromising neutrophil viability. Strikingly, better killing was achieved, while dampening neutrophil ROS responses, arguing for a more directed antifungal response and less potential to cause collateral damage85. Enhanced C. albicans clearance may stem from improved recognition, enabling more effective neutrophil responses, such as more directed NET formation or enhanced phagocytosis. While C. albicans masks β-glucans with mannoproteins to evade detection, S. cerevisiae exposes β-glucans, activating Dectin-1-mediated neutrophil responses and countering immune evasion, enhancing neutrophil phagocytosis and antifungal activity86–89. Although S. cerevisiae itself was not killed at significant extent, and does not appear to act as a harmful distraction for immune cells. Instead, its exposed β-glucans may prime or potentiate antifungal responses. Additionally, C. albicans metabolic stress in response to S. cerevisiae may increase vulnerability to oxidative stress90. By inhibiting filamentation, S. cerevisiae also enables neutrophils to prioritize phagocytosis and intracellular killing over tissue-damaging antimicrobial mechanisms like degranulation and oxidative burst91,92. Additionally, S. cerevisiae treatment enhanced neutrophil-mediated C. albicans biofilm clearance. While the underlying mechanism remains to be fully elucidated, it may involve disruption of biofilm-associated barriers, including extracellular matrix density, hyphal architecture, and altered PAMP expression, that typically impair immune recognition and contribute to fungal persistence93. In vivo, Sc3458 treatment significantly reduced C. albicans burden. However, it is challenging to attribute this reduced fungal burden to increased neutrophil functionality due to the difficulty of quantifying in vivo neutrophil-mediated killing. Our in vitro findings support the notion that Sc3458 not only modulates neutrophil activation but also enhances their antifungal function, which may have contributed to reduced fungal burden in vivo. While the industrial control strain also improved neutrophil-mediated C. albicans killing in vitro, our in vivo findings revealed no effects on fungal burden, accompanied by sustained pro-inflammatory responses. These contrasting outcomes underscore the strain-specific properties of S. cerevisiae in vivo. Additionally, given the physiological and microbiome differences between mice and humans, further studies, particularly clinical studies, are needed to validate these findings and elucidate the underlying pathways driving neutrophil modulation. Our screening of S. cerevisiae isolates sets Sc3458 apart from other strains as a uniquely effective candidate therapeutic agent. We highlight the potential of Sc3458 as a dual-action VVC treatment, uniquely targeting both C. albicans pathogenicity and modulating immune responses (Fig. 9). By reducing fungal load, S. cerevisiae alleviates epithelial damage, dampens pro-inflammatory axes, prevents excessive neutrophil activation, and thereby symptomatic inflammation. Simultaneously, it recalibrates neutrophil antifungal mechanisms, enhancing fungal clearance without reliance on tissue-damaging responses such as excessive ROS production. Collectively, this translated to improved fungal clearance and a shift in the immune response away from IL-1-driven and allergic hyperinflammation upon treatment of mice with Sc3458. These findings not only advance our understanding of fungal-host-pathogen interactions but also pave the way for clinical studies to validate S. cerevisiae as a live biotherapeutic agent to manage fungal infection and inflammatory responses in VVC.Fig. 9Schematic overview of S. cerevisiae-mediated interference with pathogenicity and inflammation during VVC.Illustration of the dual therapeutic potential of S. cerevisiae (blue cells) in mitigating VVC. On the left, C. albicans-driven (beige cells) VVC is characterized by a high fungal burden, activation of epithelial signaling pathways (PI3K/Akt, MAPK, NF-κB), and recruitment of dysfunctional, anergic neutrophils (neutrophils; green), leading to excessive inflammation and epithelial damage. On the right, S. cerevisiae treatment reduces fungal load, prevents overactivation of epithelial signaling pathways, and attenuates inflammatory responses. Simultaneously, S. cerevisiae enhances neutrophil antifungal activity, promoting efficient fungal clearance while minimizing tissue damage. This S. cerevisiae-mediated modulation of host immunity and fungal virulence disrupts the pathological cycle of inflammation and symptoms associated with VVC. “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. Schematic overview of S. cerevisiae-mediated interference with pathogenicity and inflammation during VVC. Illustration of the dual therapeutic potential of S. cerevisiae (blue cells) in mitigating VVC. On the left, C. albicans-driven (beige cells) VVC is characterized by a high fungal burden, activation of epithelial signaling pathways (PI3K/Akt, MAPK, NF-κB), and recruitment of dysfunctional, anergic neutrophils (neutrophils; green), leading to excessive inflammation and epithelial damage. On the right, S. cerevisiae treatment reduces fungal load, prevents overactivation of epithelial signaling pathways, and attenuates inflammatory responses. Simultaneously, S. cerevisiae enhances neutrophil antifungal activity, promoting efficient fungal clearance while minimizing tissue damage. This S. cerevisiae-mediated modulation of host immunity and fungal virulence disrupts the pathological cycle of inflammation and symptoms associated with VVC. “Created in BioRender. Van Dijck, P. (2026) https://BioRender.com/5dxv3hu”. Methods The research presented in this manuscript complies with all relevant ethical regulations. Buffy coats or venous blood samples were obtained from healthy volunteers who gave written informed consent in compliance with the declaration of Helsinki. The Jena institutional ethics committee approved the procedure (Ethik-Komission des Universitätsklinikums Jena, Permission no. 2207-01/08). The animal experimental protocols were reviewed and approved by the Ethical Committee for Animal Experimentation of KU Leuven under approval number Bio-157/2024. C. albicans and S. cerevisiae strains and culture conditions During this research, the strains utilized included C. albicans SC531494, CA315395, and I49 (vaginal isolate). For the S. cerevisiae strains, various S. cerevisiae strains were employed, namely 4609 (PYCC), 6883 (DBVPG), 6886 (DBVPG), 6888 (DBVPG), 6894 (DBVPG), 3458 (NCYC), 3471 (NCYC), 3076 (NCYC), Y.00240 (NCAIM), S288C (Cold Spring Harbor Laboratory), and CNCM I-3856 (Cerevisia®). Notably, the CNCM I-3856 strain was isolated directly from the commercial probiotic product. Before conducting all assays, individual colonies were selected from yeast extract peptone dextrose (YPD) agar plates and transferred to liquid YPD medium. Subsequently, the cultures were incubated overnight in a shaking incubator (240 rpm) at 30 °C. The cells were harvested by centrifugation (3000 × g, 1 min), subjected to two washes in 1 × phosphate-buffered saline (PBS), and adjusted to the desired optical density. Screening of 70 S. cerevisiae strains for C. albicans virulence inhibitory properties A diverse panel of 70 S. cerevisiae strains (Supplementary Table 1) was screened in a high-throughput manner for their capacity to modulate C. albicans growth, filamentation, and epithelial adhesion. All assays were performed using fluorescently labeled C. albicans to enable accurate discrimination between fungal species in co-culture. C. albicans CA3153 was co-cultured with individual S. cerevisiae strains at an initial OD600nm ratio of 0.1 (C. albicans) to 0.3 (S. cerevisiae). Cultures were incubated for 24 h at 37 °C in VSM or RPMI 1640. Relative fungal abundances were quantified by flow cytometry (Guava EasyCyte, Luminex). To assess effects on filamentation, C. albicans was co-cultured with S. cerevisiae at an OD600nm ratio of 0.1 to 0.3 under filamentation-inducing conditions (VSM, 37 °C). Filamentation was evaluated by light microscopy at 6, 8, 24, and 48 h, and fungal morphology was classified into yeast (degree 1), pseudohyphal (degree 2), or hyphae (degree 3) to assess the degree of filamentation inhibition. The impact of S. cerevisiae on C. albicans epithelial adhesion was assessed using HeLa cell monolayers, seeded following methods described in Wijnants et al.96. Cells were co-inoculated with C. albicans and S. cerevisiae at an OD600nm ratio of 0.1 to 0.3 and incubated for 90 min at 37 °C. After removal of non-adherent cells, remaining fungal populations were quantified by flow cytometry. Following identification of top-performing strains, these initial high-throughput screening assays were not used in subsequent experimental stages, as they were replaced by optimized models better reflecting the vaginal niche, host-pathogen interactions, and more translational set-ups. Growth evaluation Growth was assessed by adjusting the optical densities of C. albicans and S. cerevisiae cells to 0.1 and 0.5 at 600 nm, respectively. The evaluation was conducted in VSM at a pH of 4.2 (Table S2) or in RPMI 1640 medium (Gibco)97,98. C. albicans growth in the co-cultures was assessed at 37 °C and 760 rpm in 96-well plates, after 6 and 24 h, using fluorescently tagged C. albicans strains. Fluorescence measurements (Excitation: 569 nm, Emission: 600 nm) of C. albicans, tagged using the Cip10-mScarlet-IDT plasmid99 (CaSC5314 ACT1p-mScarlet, CaCA3153A ACT1p-mScarlet, CaI49 ACT1p-mScarlet), were performed using a Synergy™ H1 microplate reader (BioTek). Imaging of cultures The imaging of growth was conducted according to the protocol used for the growth evaluation assay. Cocultures were visualized using a Fluoview FV1000 confocal laser scanning microscope, with the GFP fluorophore excited by a 488 nm laser and emission detected through a BA505–565 bandpass filter99. Five representative, randomly selected positions were imaged for each condition, and a representative image was selected. Culturing of vaginal epithelial cells To effectively mimic the vaginal epithelial layer, A-431 VECs (ACC 91) were employed. These cells, derived from vulva epidermoid carcinoma, are frequently utilized as a representative model for the vaginal niche100,101. Before experimentation, the cells were screened for mycoplasma contamination using a PCR mycoplasma test kit (Sigma-Aldrich). A-431 VECs were cultured in RPMI 1640 medium (Gibco), supplemented with 10% fetal calf serum (FCS), and maintained at 37 °C in a 5% CO2 atmosphere with constant humidity. During experimental assays, cells were seeded at the desired density. Adhesion to VECs VECs were seeded at a density of 2 × 104 cells per well in 96-well plates and incubated at 37 °C with 5% CO2 for two days, or until confluent. C. albicans and S. cerevisiae cells were both adjusted to a Multiplicity of Infection (MOI) of 5. VECs were infected with C. albicans cells, with or without an S. cerevisiae strain (100 µL of 106 cells/mL in RPMI 1640 without FCS) and incubated for 90 min at 37 °C and 5% CO2. In the assessment of S. cerevisiae’s potential to displace C. albicans, C. albicans was initially introduced alone for 90 min. Subsequently, non-adherent cells were removed by rinsing with PBS, and the S. cerevisiae cells were added for an additional 90 min. For the final readout, the medium was removed, and the wells were gently rinsed with PBS to eliminate non-adherent or loosely adherent cells. To detach the adhered C. albicans cells, the wells underwent trypsinization using 0.5% trypsin EDTA (Gibco), followed by sonication and dilution for a suitable colony forming units (CFU) readout on CHROMagarTM Candida (BioTrading), enabling differentiation between C. albicans and S. cerevisiae cells. Filamentation evaluation C. albicans filament formation was assessed in both VSM and RPMI 1640. The experimental setup was similar to that of the growth assay, except for adjusting the VSM pH to 5 to facilitate early filament formation. Following a 3-h incubation at 37 °C at 400 rpm, filament formation was quantified by manually counting filamentous cells using a Bürker chamber under a Zeiss microscope. Epithelial damage to VECs VECs were seeded as described in the adhesion assay and infected with C. albicans and S. cerevisiae cells as described above, however, both at MOI 0.5, and incubated for 24 h at 37 °C with 5% CO2. To assess necrotic epithelial damage, the release of the cytoplasmic enzyme lactate dehydrogenase (LDH) was measured using the Cytotoxicity Detection Kit from Roche102. LDH activity was quantified in the supernatant of infected VEC monolayers in 96-well plates 24 hpi. LDH from rabbit muscle (5 mg/ml, Roche) was utilized to establish a standard curve for determining LDH concentrations. The background LDH activity from uninfected VECs was subtracted from the test conditions. Cytokine release by VECs Similar to the setup above, supernatants from VECs infected with C. albicans, with or without the addition of S. cerevisiae, were analyzed for the presence of inflammatory cytokines (IL-1α, IL-8, GM-CSF). Supernatant cytokine concentrations were measured using human enzyme-linked immunosorbent assays (ELISA; R&D Systems) following the manufacturer’s instructions. Cluster analysis Cluster analysis was performed to identify patterns of similarity and the top-performing S. cerevisiae strains based on their rankings across the assays. Clustering quality was assessed using the silhouette width metric, which ranges from 1 (indicating a strong fit) to −1 (indicating a poor fit). To determine the optimal number of clusters (k), K-means and K-medoids algorithms were applied, with the best clustering identified by maximizing the silhouette width103. Data analysis and visualization were performed using RStudio (RStudio Team, 2020) with the R programming language. Hemolysis assay Hemolytic activity of C. albicans and S. cerevisiae strains was assessed using Sabouraud Dextrose Agar supplemented with 5% defibrinated sheep blood (ThermoFisher Scientific). Yeast suspensions were adjusted to 1 × 10⁷ cells/mL, and 10 µL was spotted onto blood agar plates. Plates were incubated for 24 h and 48 h at 37 °C with 5% CO2. Hemolysis was evaluated based on the presence and type of hemolytic activity surrounding the colonies. β-hemolysis was identified by the formation of a clear zone around colonies. For quantitative assessment, both colony diameter and the diameter of the surrounding hemolytic halo were measured. The β-hemolytic index was calculated as the ratio of halo diameter to colony diameter. Acid tolerance Acid tolerance was assessed by evaluating yeast growth based on optical density measurements, as described for the growth evaluation assay. C. albicans and S. cerevisiae cells were adjusted to an OD₆₀₀ of 0.5 and incubated in vaginal simulation medium (VSM) adjusted to pH 3.4, 3.6, 3.8, 4.0, 4.2, 4.4, or 4.6. Growth assays were performed in 96-well plates at 37 °C with shaking at 760 rpm for 24 h. Optical density was measured using a Synergy™ H1 microplate reader (BioTek). RNA isolation and pooling VEC were seeded in 6-well plates at 3 × 105 cells per well and cultured until confluent for 2 days. Following medium removal, VECs were infected with C. albicans at MOI 5, either alone or with S. cerevisiae at MOI 5 (750 µL of 3 × 106 cells/mL in RPMI 1640 without FCS) at 37 °C with 5% CO2. RNA samples were collected at 6 and 24 hpi. Plates were centrifuged cold for 5 min at 500 × g, after which the supernatant was replaced with 500 μL of RLT buffer (QIAGEN), containing 1% β-mercaptoethanol. The cells were detached using a cell scraper, immediately shock-frozen in liquid nitrogen, and stored at −80 °C until further use. For RNA extraction, the samples were thawed on ice and centrifuged for 10 min at 15,000 × g at 4 °C. The supernatant was collected and transferred to a new microcentrifuge tube for the isolation of human RNA, while the resultant pellet was used to isolate fungal RNA. Both human and fungal RNA were isolated using the TRIzol-chloroform method, as previously described104. The RNA concentration and quality of both human and fungal RNA were assessed using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific). To ensure sufficient library depth for both human and fungal RNA, we evaluated various pooling strategies105. Based on these evaluations, fungal and human RNA samples were combined in a 1:2 ratio by concentration for library preparation and sequencing, aiming to achieve adequate sequencing depth for comprehensive differential gene expression analysis. Before pooling, we measured cross-mapping rates by aligning reads from strictly human and fungal samples to respective fungal and human genomes, resulting in no cross-mapped reads between the three species. RNA-seq library preparation and sequencing Bulk RNA sequencing (RNA-seq) libraries were prepared by Novogene (Directional mRNA enrichment libraries) and subjected to NovaSeq paired-end sequencing (2 × 150 base pairs). Sequencing depth and read length were optimized to ensure comprehensive coverage and reliable differential gene expression analysis. RNA-seq pipeline Paired-end FASTQ files were obtained from Novogene for subsequent analysis. Initial quality control of the raw sequencing data was conducted using FastQC and MultiQC106,107. For read alignment and quantification, we employed the splice junction-sensitive aligner STAR108. Given that the samples included RNA from both host and pathogen, we performed read mapping against a concatenated reference comprising both the human and C. albicans genomes. Human reads were aligned to the primary genome assembly GRCh38, with annotations derived from NCBI (last accessed May 2, 2024). The C. albicans SC5314 genome assembly ASM18296v3 and associated annotations were sourced from NCBI (last accessed May 2, 2024). BEDtools (intersectBed) and Picard were used to remove rRNA reads and optical duplicates from the mapped reads109,110. Samtools was used for indexing and generating summary statistics111. htseq-count was used to count the number of genes within each BAM file112. Coverage was evaluated using deepTools (bamCoverage) and BEDtools110,113. Sequence alignment was assessed using BLAST114. Data analysis and visualization were performed using RStudio (RStudio Team, 2020) with the R programming language. Differential gene expression analysis was performed using the Bioconductor package edgeR v.3.34.0115. Read counts were obtained from STAR mapping. For human samples and each fungal species, we compared time point 0 with other time points throughout the infection course using the glmQLFTest function from edgeR. We used filterByExpr to filter out low-count genes. The dispersions were estimated using the estimateDisp function, and differential expression was assessed by applying a contrast matrix with glmQLFTest. Genes with a log2(fold change) > 1.0 and an adjusted P (FDR) < 0.05 were considered differentially expressed. To visualize the data, we performed Principal Component Analysis (PCA) on normalized counts and created plots using ggbiplot. Differentially expressed gene results were visualized with scatter and volcano plots. Gene ontology enrichment was performed using ShinyGO116. Following this, the significantly enriched GO terms were analyzed using REVIGO85, applying a similarity threshold of Small (0.5)117. Biofilm biomass C. albicans biomass formation on VECs was assessed by measuring the fluorescence emitted by fluorescently tagged C. albicans strains. The experimental procedure followed the adhesion assay protocol: VECs were seeded and cultured until confluent, then infected with C. albicans, with or without S. cerevisiae. At 24 and 48 hpi, the medium was aspirated, and the wells were gently washed with PBS to remove non-adherent or loosely attached cells. Subsequently, biomass quantification was performed using fluorescence measurements on a Synergy™ H1 microplate reader (BioTek), with excitation set to 569 nm and emission to 590 nm99. For experiments involving transwells, the cells were physically separated using a ThinCertTM Cell Culture insert (Greiner Bio-one) with a 0.4 µm pore size. For experiments involving heat-inactivated S. cerevisiae cells, inactivation was performed by incubating the cells in 1X PBS at 65 °C for 30 min. Biofilm distribution Biofilms were established using the same protocol as the biofilm biomass assay. At 24 hpi, the supernatant was collected, and the wells were gently washed with PBS to remove non-adherent or loosely attached cells. To detach the adhered C. albicans cells, the wells were treated with 0.5% trypsin-EDTA (Gibco), followed by water bath sonication. The resulting cell suspension was then diluted to obtain an appropriate CFU count on CHROMagar™ Candida (BioTrading). Biofilm structure Biofilms were established following the same protocol as the biofilm biomass assay but were grown on IBIDI plates. Imaging of the biofilms was performed using a Fluoview FV1000 confocal laser scanning microscope. The mScarlet fluorophore was excited with a 559 nm laser, and emission was captured through a BA575–675 bandpass filter99. A series of images were acquired at 4.51 μm intervals along the z-axis, allowing for a three-dimensional reconstruction of the biofilms from the start to the end of the fluorescent signal. The z-stack images were subsequently exported to the Napari interface for further analysis118. Preparation of hMDMs Buffy coats obtained from healthy volunteers underwent Histopaque-1077 (Sigma-Aldrich) density centrifugation to isolate human peripheral blood mononuclear cells (hPBMC). CD14-positive monocytes were subsequently isolated using magnetic automated cell sorting (autoMACs; MiltenyiBiotec). These cells were then differentiated into human monocyte-derived macrophages (hMDMs) by seeding 1 × 107 cells in a 175 cm2 cell culture flask containing RPMI 1640 medium with 2 mM l-glutamine (Thermo Fisher Scientific), supplemented with 10% heat-inactivated fetal bovine serum (FBS; Bio&SELL), and 50 ng/mL recombinant human M-CSF (ImmunoTools). The cells were incubated for seven days at 37 °C and 5% CO2 with a medium exchange on day 5. The adherent hMDMs were detached using 10 mM EDTA in PBS, then seeded in 96-well plates at a final concentration of 4 × 104 hMDMs/well in RPMI 1640 + FBS + M-CSF, and incubated overnight. The macrophage infection experiments were conducted in serum-free RPMI 1640 medium. hMDMs stimulation To assess the stimulation of hMDMs, the release of IL-1β was measured, as it serves as an indicator of NLRP3 activation41. hMDMs were seeded at a density of 4 × 104 cells per well, as previously described. Before infection, the phagocytes were primed by incubating them with 50 ng/mL of lipopolysaccharide (LPS) (Sigma-Aldrich) for 2 h. Subsequently, the hMDMs were infected with C. albicans at MOI 0.5, with or without S. cerevisiae at MOI 0.5, for 24 h at 37 °C with 5% CO2. After the designated incubation period, the supernatant from the infected hMDMs was collected by centrifuging the plates at 250 × g for 5 min. Supernatant IL-1β concentrations were measured using human enzyme-linked immunosorbent assays (ELISA; R&D Systems) following the manufacturer’s instructions. Isolation of human neutrophils The isolation of primary human neutrophils followed the procedures outlined by Gresnigt et al.119. Initially, human peripheral blood mononuclear cells were separated from granulocytes and erythrocytes using density gradient centrifugation with Histopaque-1077 (Sigma-Aldrich) in a sterile 50 mL falcon. Neutrophils were subsequently extracted from the erythrocyte/granulocyte fraction by hypotonic lysis with a buffer containing 155 mM NH4Cl and 10 mM KHCO3. Following two washes in 1 × PBS, the neutrophils were resuspended in RPMI 1640 medium and adapted to the required concentration. Neutrophil stimulation: IL-8 release To assess neutrophil stimulation in response to C. albicans and the potential influence of S. cerevisiae, neutrophils were treated with supernatants. These supernatants were diluted 1:2 in RPMI 1640 and collected from VECs exposed to C. albicans, with or without S. cerevisiae, for 24 h. This process was carried out as previously described. Freshly isolated neutrophils were seeded at 1 × 105 neutrophils/well in a 96-well plate and following a 3-h stimulation period at 37 °C with 5% CO2, supernatants from the neutrophils were collected. The concentration of IL-8, an indicator of neutrophil activation, was then measured in these supernatants using an ELISA assay (R&D Systems) following the manufacturer’s instructions120. Neutrophil activation: ROS production Neutrophil activation in response to C. albicans, with and without S. cerevisiae, was evaluated using ROS production in an oxidative burst assay. To quantify the total ROS, a luminol-enhanced chemiluminescence method was performed. Freshly isolated neutrophils were seeded at 1 × 105 neutrophils/well as mentioned above. Neutrophils were then incubated for 30 min at 37 °C and 5% CO2, to let them attach. Subsequently, the neutrophils were stimulated for 1 h 37 °C and 5% CO2 with undiluted supernatants from VECs exposed to C. albicans, with and without S. cerevisiae following the previously described protocol. After incubation, 50 µL of zymosan (at a concentration of 150 µg/mL) was added to further stimulate the fungal immune response and enhance ROS production121. Following stimulation, 50 µL of RPMI 1640 containing 200 µM luminol (Sigma-Aldrich) and 16 U HRP (Sigma-Aldrich) for total ROS detection was added. Chemiluminescence was measured every 2.5 min for 2.5 h using a Tecan Infinite M200 microplate reader. The area under the curve was calculated using GraphPad Prism 10.4.3. Neutrophil stimulation: activation-associated cell surface markers Further assessment of neutrophil stimulation involved analyzing the regulation of neutrophil activation markers using flow cytometry. The selection of activation markers was based on granulocyte responses to fungal pathogens or related stimuli40,42. Neutrophils were seeded in a round-bottom 96-well plate at a density of 2 × 105 cells per well and stimulated with undiluted supernatants from VECs exposed 24 h to C. albicans, with and without S. cerevisiae, following the previously described protocol. After a 3 h stimulation period at 37 °C with 5% CO2, supernatants were removed, and the neutrophils were washed with flow cytometry buffer (PBS, 2% FCS). To minimize unspecific staining, neutrophils were first incubated with Fc-block Human TruStain FcX (BioLegend). Subsequently, a mixture of fluorophore-linked antibodies targeting surface molecules indicative of activation status was added: CD15-APC-Fire750 (W6D3), CD11b-BV421 (ICR44), CD16-PerCP-Cy5.5 (3G8), CD35-FITC (E11), CD62L-AlexaFluor647 (DREG-56), CD66b-PE (G10F5), and CXCR2 (5E8, all from BioLegend). Fixable Viability Dye eFluor506 (Invitrogen) was added to exclude dead cells for 20 min at 8 °C. Subsequently, cells were washed with flow cytometry buffer, filtered through a 70 µm mesh, and then analyzed using a FACSVerse Cell Analyzer flow cytometer (BD Biosciences). Gating strategies for granulocyte populations and activation markers followed the approach described by Valentine et al.40. Data analysis was carried out using FlowJo v.10. Neutrophil chemotaxis Neutrophil chemotaxis was performed as described by Valentine et al.40. Briefly, primary human neutrophils were stained with Cytopainter Green (2 µL in 1 × 106 neutrophils suspended in RPMI 1640), incubated for 10 min at room temperature in the dark. Cells were then washed with Hank’s Balanced Salt Solution containing 20 mM HEPES (pH 7), centrifuged (10 min, 300 × g), and resuspended in endothelial cell medium (ECM, PromoCell) at 5 × 105 cells/mL. Human umbilical vein endothelial cells (HUVECs) were cultured for 72 h in 150-cm2 flasks, harvested, and seeded into 3 µm transwell inserts at 2 × 104 cells/insert. After 48 h, inserts were transferred to 24-well plates containing confluent VECs, seeded two days earlier at 1 × 105 cells/well. Medium in the inserts was refreshed (200 µL), and VECs were infected with C. albicans (1 × 105 cells, MOI 0.5) in the presence or absence of S. cerevisiae (1 × 105 cells, MOI 0.5) in a total volume of 600 µL. After 18 h, 200 µL of stained neutrophils were added to the inserts. Plates were incubated for 2 h at 37 °C and 5% CO₂. Images were acquired using an IncuCyte SX5 live-cell imaging system (Sartorius). Thresholds for fluorescence intensity and area size ensured accurate segmentation and quantification. Image analysis was performed using Incucyte 2023 A Rev2 software. Neutrophil death and NETosis evaluation To assess the influence of S. cerevisiae on C. albicans-induced neutrophil responses, co-cultures were established to evaluate NETosis and neutrophil cell death. Neutrophils (1 × 10⁵ cells per well) were pre-incubated with Sytox Green (4 × 10⁻⁵ µL/mL RPMI 1640) to monitor NETosis and cell death, while S. cerevisiae and C. albicans were added at an MOI of 0.5. Preincubation was performed in the dark at 37 °C with 5% CO₂ for 15 min prior to co-culture. Co-cultures were imaged in an Incucyte SX5 (Sartorius) at 37 °C with 5% CO₂. Negative controls were left unstimulated, positive controls were stimulated with 20 nM PMA. NETosis and cell death were monitored in real-time using green fluorescence imaging. Images were captured from two regions per well every 20–30 min for 12 h at ×20 magnification, with an exposure time of 100 ms. Representative images and images of unstimulated and PMA-induced NETosis conditions were used to train the Incucyte Basic Software for analysis. The Top-Hat method was applied for background correction, and the edge-split tool was used to distinguish closely spaced objects. Thresholds for fluorescence intensity and area size ensured accurate segmentation and quantification. Image analysis was performed using Incucyte 2023A Rev2 software. Neutrophil-mediated C. albicans killing and growth To evaluate the effects of S. cerevisiae on neutrophil-mediated C. albicans growth impairment, we used a setup similar to that described above, incorporating mScarlet-tagged C. albicans. Fungal overgrowth was monitored in the orange fluorescence channel using the Incucyte live-cell imaging system, with an exposure time of 300 ms. The Adaptive method was applied for image analysis to account for variations in fluorescence intensity and object morphology during fungal growth. Data were analyzed using Incucyte 2023A Rev2 software. To assess the influence of S. cerevisiae on neutrophil-mediated C. albicans killing, a similar co-culture system was used, except that unstained neutrophils and C. albicans were included. Killing assays were performed at 1 and 3 hpi. After co-culture, neutrophils were lysed with 4% Triton X-100, and samples were scraped, sonicated, serially diluted, and plated on CHROMagar™ Candida (BioTrading) to differentiate C. albicans from S. cerevisiae. Neutrophil-mediated C. albicans biofilm eradication To evaluate whether the biofilm-inhibiting effects of S. cerevisiae enhance C. albicans clearance by neutrophils, we assessed neutrophil-mediated killing in co-culture with C. albicans biofilms. Following the biofilm biomass quantification protocol, co-cultures were incubated for 24 h, washed to remove non-adherent cells, and fluorescent C. albicans values were measured to normalize the data, accounting for reduced biofilm biomass due to S. cerevisiae. 5 × 10⁵ neutrophils per well were added and incubated with the biofilms for 3 h at 37 °C with 5% CO₂ to allow for C. albicans killing. After co-culture, neutrophils were lysed with 4% Triton X-100, and biofilms were scraped, sonicated, serially diluted, and plated on CHROMagar™ Candida (BioTrading) to differentiate C. albicans colonies from S. cerevisiae. This method enabled precise quantification of C. albicans survival and biofilm clearance. Murine model of VVC Female BALB/cJ mice 8 weeks old were obtained from Charles River and used for the study. Mice were housed with a 14 h light and 10 h dark cycle between 22–24 °C and at 45–70% humidity. To induce a pseudoestrous state, mice received a subcutaneous injection of 0.5 mg β-estradiol 17-valerate (Sigma), dissolved in 100 µL of sesame oil, 5 days before inoculation. Estrogen administration was continued weekly throughout the experiment72. To establish infection, estrogen-treated mice were sedated and inoculated with 10 µL of 107 bioluminescent (BLI) C. albicans CA3153 or SC5314 cells/mL, suspended in 1 × PBS. The inoculum was administered intravaginally via pipetting. To minimize leakage and enhance fungal retention, mice were positioned with their hindquarters elevated for 20 min post-inoculation. Following a 24-h incubation period to allow for infection establishment, mice received intravaginal treatment with either 10 µL of 108 S. cerevisiae 3458 cells/mL or 1 × PBS (control), similar to C. albicans infection. Treatments were initiated one day post-inoculation and continued daily. Monitoring C. albicans vaginal load Fungal burden was assessed on days 2, 4, 6, 8, and 10 post-infection using bioluminescence imaging (BLI). Mice were anesthetized with 2.5–3.5% (v/v) isoflurane gas, followed by maintenance at 1.5–2% isoflurane in oxygen (Piramal Critical Care). D-luciferin (1.1665 mg/mL, 10 µL) was administered intravaginally. After anesthesia induction, mice were positioned supine in an IVIS Spectrum System (Perkin-Elmer, Hopkinton, MA, USA). Imaging began 5 min post-injection, capturing five consecutive images using the following parameters: 30-s exposure time, medium binning, F/stop of 1, and subject height of 1.5 cm. Bioluminescence signal quantification was performed using Living Image Software (version 4.7.3, Perkin-Elmer). Total photon flux (photons/second) was measured within a region of interest (ROI) of 2.8 cm2, covering the vaginal area. The maximum signal from consecutive acquisitions was used for further analysis. Monitoring inflammatory state To assess the presence of inflammatory markers, vaginal lavages were performed. Each lavage was conducted using 200 µL of 1 × PBS, administered in four consecutive volumes of 50 µL each. Following collection, the lavage samples were centrifuged at 200 × g for 10 min at 4 °C to separate soluble components from immune cells. The resulting supernatants were collected and subjected to quantitative analysis using a Luminex Mouse Discovery Assay. Flow cytometry Flow cytometry was performed to assess the presence of macrophages, neutrophils, and eosinophils in vaginal lavage fluid. Lavage samples were centrifuged at 200 × g for 10 min at 4 °C to separate soluble components and immune cells. The resulting cell pellet was washed with ice-cold 1 × PBS and incubated with Fc-receptor blocking antibodies (anti-CD16/CD32 Miltenyi Biotec, Bergisch Gladbach, Germany) for 15 min at room temperature in the dark (Table S3). Simultaneously, a Zombie Aqua™ viability dye (BioLegend, USA) was applied to discriminate live from dead cells. Following incubation, the cells were washed with cold 1 × PBS supplemented with 2% FCS-2mM EDTA and stained with a surface marker antibody cocktail (Table S3) at 4 °C for 30 min. After staining, cells were washed with cold PBS containing 2% FBS and resuspended in PBS supplemented with 2% v/v FBS and 2 mM EDTA to prevent cell aggregation. Samples were maintained at 4 °C until analysis. Flow cytometry was conducted using a BD LSRFortessa™ X-20 (BD Biosciences) and data analysis was carried out using FlowJo software (BD Biosciences, v10). The specific gating strategies are depicted in Fig. S16. Vaginal histology Vaginal tissues were fixed in 4% paraformaldehyde and embedded in paraffin. Paraffin-embedded sections were cut at a thickness of 5 μm from the mid-vaginal region. Sections were subsequently deparaffinized and subjected to histological staining with hematoxylin and eosin122. Histology was scored for inflammation by a blinded experimenter as adapted from ref. 123. Histological sections of proximal, mid, and distal regions were imaged at 20× magnification and at least 10 views were inspected, with the score averaged over these views. Score 0 was given upon “no Inflammation”: no neutrophils observed in the epithelial layer and the vaginal lumen. 1 “Mild inflammation”: some neutrophils observed in the lumen and some infiltrated in the epithelial layer. 2 “Moderate inflammation”: neutrophils or small microabscesses lining the top epithelial layers. 3 “Severe inflammation”: substantial neutrophils and/or larger microabscesses observed in the mucosa layer yet still regions without inflammation observed. 4 “very severe inflammation”: substantial neutrophils and/or larger (ruptured) microabscesses in the mucosa and visual distortion of mucosal barrier and mucosal regions without inflammatory infiltrates are scarce. Statistics and reproducibility Data from all assays except the RNA-sequencing different gene analysis were analyzed using GraphPad Prism version 10 (GraphPad Software, La Jolla, California, USA). Data from at least three biological replicates were analyzed and values are presented as mean ± standard error of the mean (SEM). All microscopy analysis where at least independently repeated 2 times and representative images are provided in the manuscript. Used statistical tests are indicated in the figure legends. Statistical significance is depicted in the figures as P values. All data used to generate the graphs are provided as source data. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting Summary Transparent Peer Review file Supplementary Information Reporting Summary Transparent Peer Review file Source data Source data Source data