Local albumin excess exacerbates Candida albicans-induced inflammasome activation linked with hyperinflammation during vulvovaginal candidiasis

Abstract

Vulvovaginal candidiasis (VVC) is a mucosal yeast infection, mostly caused by Candida albicans, where symptoms are driven by inflammation. The onset of VVC is instigated by C. albicans pathogenicity, but the underlying causes of disease-driving hyperinflammation remain incompletely resolved. We found that vaginal albumin concentrations are higher in women with recurrent VVC and during experimental VVC in mice. These albumin levels correlated with inflammatory cytokines in vaginal lavages. While the abundance of this serum protein in the vagina can be a hallmark of inflammation-induced vascular permeability, we reveal molecular mechanisms by which albumin can also drive inflammation. Albumin increased NLRP3 inflammasome activation in human macrophages elicited by C. albicans. This was associated with increased resistance of C. albicans to macrophage-mediated clearance in the presence of albumin. Further, albumin increased the expression of fungal secreted aspartic protease 1, which we identified as the effector driving inflammasome activation and IL-1β responses in vitro and in an in vivo VVC model. Moreover, neutrophils also exhibited impaired clearance of C. albicans in the presence of albumin. Collectively, vaginal albumin may not only be a consequence of inflammation of the vaginal mucosa, but also can drive C. albicans-induced hyperinflammatory IL-1β responses that underlie immunopathology in VVC.

This study shows that high vaginal albumin in vulvovaginal candidiasis is not merely a correlate of inflammation, but can drive dysfunctional hyperinflammation by promoting fungal induced inflammasome activation, while compromising fungal clearance.

Introduction

Vulvovaginal candidiasis (VVC) is one of the most common infections of the female reproductive tract, occurring in 70–75% of otherwise healthy women at least once during their life¹. Particularly for women experiencing several episodes per year, referred to as recurrent VVC (RVVC), quality of life is severely affected^2,3. VVC symptoms are driven by an inflammatory response against the mucosal infection with Candida species⁴, most commonly C. albicans⁵. This inflammatory response is instigated by fungal pathogenicity, though its persistence and development of symptoms result from neutrophil responses, in part driven by the NLRP3 inflammasome^4,6–8. The recruited neutrophils to the vulvovaginal mucosa during VVC are dysfunctional in fungal clearance and cause immunopathology^6,8.

The pathogenicity of C. albicans in the context of VVC involves several processes. The toxin candidalysin causes epithelial damage and triggers immune responses in the vaginal mucosa that drive downstream inflammation^9–11. The trace mineral binding protein Pra1 helps C. albicans to overcome zinc starvation, yet can be recognized by the immune system to elicit neutrophil recruitment¹². Other infection-specific proteins involved during VVC are C. albicans secreted aspartic proteinases (Saps)¹³. Some of these virulence factors trigger the activation of the NLRP3 inflammasome^14,15, a multiprotein complex that is highly expressed in macrophages¹⁶, which are abundant immune cells in both the human and murine vaginal mucosa^17–20. The NLR family pyrin domain containing 3 (NLRP3) inflammasome processes pro-interleukin (IL)−1β into its bioactive proinflammatory form²¹, which subsequently is a strong driver of neutrophil-mediated inflammation^7,22,23. While the association between NLRP3 inflammasome activation and VVC is clear, it is not completely understood what drives its characteristic hyperactivation in this disease.

VVC mainly affects women in their fertile years²⁴. Among other predisposing factors, the use of hormonal replacement therapy, pregnancy, the use of oral contraceptives, and antibiotics are associated with VVC^1,24,25. Estradiol is the predominant form of estrogen in the female body and its concentration depends on the menstrual cycle²⁶, and exogenous administration. To mimic and study VVC in mice, estradiol administration is required to successfully establish C. albicans infections in the vaginal tract. Collectively, this highlights the central role of estradiol in the susceptibility to VVC. Besides its major role in reproductive health, estradiol increases the vascular permeability of the female reproductive organs²⁷. Acute local inflammatory responses can further increase the vascular permeability. Consequently, albumin, as the most abundant serum protein, is prevalently found in the vaginal fluid and cervical secretions, and significantly increases particularly in the context of infection and inflammation^28,29 as well as intra uterine device use³⁰. We previously identified that albumin induces adaptations in C. albicans and C. glabrata that can enhance their pathogenicity towards epithelial cells^31,32. Yet, albumin can also neutralize candidalysin³³, a key driver of VVC-associated inflammation and inflammasome activation^7,15,22,23. Thus, while albumin is present in the vagina during inflammation, its role in the inflammatory pathogenesis of VVC remains unclear.

Acute inflammation driving vascular permeability can cause albumin influx into tissues. Accordingly, vaginal albumin levels correlate with inflammation during experimental VVC in mice and in patients with RVVC. Here, we show that albumin also can serve as a mechanistic driver of C. albicans-induced hyperinflammatory responses. Albumin induced C. albicans pathogenicity mechanisms enabling it to resist macrophage-mediated killing and elicit stronger NLRP3 inflammasome activation.

Results

Vaginal albumin levels are linked to inflammation during VVC

To define the association of vaginal albumin with VVC, albumin was quantified in vaginal lavages from healthy women and RVVC patients of a previously collected cohort³⁴. In line with previous observations²⁸, relative vaginal albumin abundance was elevated in RVVC patients compared to healthy women (Fig. 1A). To understand how albumin is linked with inflammatory responses during RVVC, a Pearson correlation analysis was performed. Significant correlations between albumin concentrations and the proinflammatory cytokines IL-1β, IL-6, and IL-17 and the neutrophil chemoattractant IL-8 were found in vaginal lavages from RVVC patients (RVVC₁), but not in healthy women (Fig. 1B, C and Supplementary Fig. S1). Samples from an independent multi-center RVVC patient cohort (RVVC₂)^35,36 similarly showed significant correlations of IL-1β, IL-6, and IL-8 with the abundance of albumin. Subtle correlations with albumin were also found for vaginal pH, pregnancy, number of children, hay fever, and disturbance of lactobacilli-dominated vaginal flora (all positive correlations) and duration of symptoms (negative correlation) (Supplementary Fig. S2). Multivariate Principle Component analysis shows that the direction of variance in the patients driven by albumin clusters in a similar direction as cytokines, leukocytosis, vaginal pH, and lactobacilliary grade (Supplementary Fig. S3). Additionally, IL-1 receptor antagonist (Ra) and TNF positively correlated to albumin concentrations in the vaginal lavages collected within the RVVC₂ cohort (Fig. 1B and Supplementary Fig. S1).

Fig. 1. Albumin concentrations correlate with VVC-associated cytokines in vivo.

A Albumin concentrations in vaginal lavages of healthy women (n = 9) and RVVC₁ patients (n = 13). Significance was tested using a two-tailed Welch’s t-test. B Heatmap showing two-tailed Pearson correlation coefficient (r) of cytokine and albumin concentrations in vaginal lavages from healthy women (n = 9) and two independent cohorts of RVVC patients (n = 13 and n = 102, respectively). C Example of two-tailed Pearson correlations shown in (B) for IL-1β and albumin in the RVVC₁ cohort and healthy women. D Mouse VVC infection model scheme. E Representative images of vaginal histology sections of naïve or C. albicans infected mice, 2 weeks after infection. Sections were stained with hematoxylin/eosin to visualize tissue pathology (H&E, scale = 100 µm), Grocott’s (Scale = 50 µm) to visualize fungal presence and immunofluorescence (scale = 50 µm) showing nuclei (DAPI, blue), C. albicans (anti-Candida antibody, green, indicated by white arrowheads) and neutrophils (anti-Ly6G, red). Images are representative of 4 mice per group. F Albumin levels in vaginal lavages (n = 13 uninfected and n = 15 infected) or organ homogenate (n = 9 uninfected and n = 12 infected) from β-estradiol-treated, and β-estradiol-treated C. albicans-infected mice. Significances were calculated using two-tailed Mann–Whitney U-test. G two-tailed Pearson correlations between albumin concentrations and colony forming units (CFUs, week 1), IL-1β concentrations, MPO concentrations, or neutrophil count (n = 9 mice) in vaginal lavages one week after infection of β-estradiol-treated mice with C. albicans. Data in (F and G) are pooled data from three and two independently conducted experiments, respectively. Bars represent the mean + SEM with dots as individual biological replicates.

To validate the link between albumin and parameters of VVC immunopathology, we performed an estradiol-based mouse VVC model (Fig. 1D). We observed establishment of fungal invasion and a strong neutrophilic cell infiltration into vaginal tissue and lumen (Fig. 1E). Relative albumin abundance significantly increased in vaginal lavages and even more pronounced in the mucosal tissue of C. albicans-infected β-estradiol-primed mice (Fig. 1F). Further linking albumin and inflammation, vaginal albumin correlated with IL-1β concentrations and myeloperoxidase (MPO) levels one week post-infection. Further, a trend (P = 0.09) towards increased neutrophil numbers in week one was observed (Fig. 1G). Strikingly, we found that albumin levels at one week post-infection significantly correlated with neutrophil numbers at two weeks post-infection (Fig. 1G and Supplementary Fig. S4), suggesting a link between albumin and inflammatory responses early during infection that drive persistent neutrophilia. Two weeks post-infection, a positive correlation between albumin and MPO release suggested a persistent link between albumin and neutrophil activation later during infection (Fig. 1G).

Albumin potentiates innate inflammatory responses

Based on the observations made in RVVC patients and the VVC mouse model, we hypothesized that albumin may not only be a consequence of inflammation but may also be a driver of inflammation. Epithelial cells mount initial inflammatory responses to vaginal C. albicans infection and recruit and activate neutrophils^7,10. Given that albumin can increase pathogenicity of Candida species towards host epithelial cells^31,32, we introduced albumin to an in vitro vulvovaginal epithelial infection model (A-431 cells)³⁷ and quantified cytokine responses. The presence of albumin did not affect C. albicans-driven IL-1α and IL-1Ra responses and even decreased IL-8 release by A-431 cells (Fig. 2A). Moreover, the cytokines IL-1β and IL-6 were not released by these cells (Fig. 2A). Even following prolonged incubation, a setting where albumin increases C. albicans-induced epithelial damage³², no substantial increase in the release of proinflammatory cytokines was observed (Supplementary Fig. S5A). Another cell line (SW954 cells) representing the vulvar epithelium, mounted modest IL-1α and IL-1β as well as IL-8 responses to C. albicans infection. Yet, the presence of albumin neutralized their release (Fig. 2B). IL-6 remained undetectable in all conditions and IL-1Ra release was not changed by albumin.

Fig. 2. Albumin drives proinflammatory cytokine release by macrophages.

A Cytokine release by A-431 cells (n = 5 repeats) in response to C. albicans infection in the presence or absence of albumin (10 mg/mL) at 24 hpi. B Cytokine release by SW954 cells (n = 4 repeats) in response to C. albicans infection in the presence or absence of albumin at 24 hpi (10 mg/mL). C Cytokine release by human monocyte-derived macrophages (n = 8 donors) in response to C. albicans in the presence or absence of albumin (10 mg/mL). D IL-1β release by human monocyte-derived macrophages in response to C. albicans in the presence or absence of albumin (5 mg/mL) and/or recombinant human IL-1Ra, anakinra (5 μg/mL; n = 7 donors). E IL-1β levels in culture supernatants of neutrophils (n = 5 donors) in response to C. albicans in the presence or absence of albumin (5 mg/mL) and/or supplemented IL-1β. Data are from at least three independently conducted experiments. Bars represent the mean + SEM with dots as individual biological replicates. Statistical analysis was done using two-way RM-ANOVA with Holm–Šídák post-hoc test.

In addition to epithelial cells, macrophages are important tissue-resident sentinels patrolling the vaginal mucosa^17–20. Interestingly, human monocyte-derived macrophages (hMDMs) mounted significantly increased cytokine responses to C. albicans in the presence of albumin (Fig. 2C). Albumin did not strongly affect their baseline viability, mitochondrial metabolism, or expression of selected surface markers indicative of polarization and  activation (Supplementary Fig. S5B–D). Enhanced C. albicans-induced IL-1β, IL-1Ra, IL-8, and IL-6 responses by hMDMs in the presence of albumin aligns with the cytokines observed to correlate with albumin in RVVC patients (Fig. 1B). IL-1β is notorious for amplifying its own release²¹, therefore we sought to dampen this using recombinant IL-1Ra (Anakinra). Macrophage stimulation in the presence of albumin and anakinra returned the IL-1β levels back to similar levels as when albumin was absent (Fig. 2D). Nevertheless, even when the self-amplifying loop is interrupted by Anakinra, IL-1β release was significantly boosted by albumin.

Following their recruitment to the vaginal mucosa, neutrophils will interact simultaneously with both albumin and C. albicans. Unlike macrophages, human primary neutrophils failed to mount detectable IL-1β responses to C. albicans, which was also not changed by the presence of albumin (Fig. 2E). In line with previous observations, when stimulated with exogenous IL-1β, neutrophils degraded the cytokine to undetectable levels³⁸. Strikingly, the presence of albumin prevented substantial degradation (Fig. 2E). Collectively, albumin enhanced inflammatory responses to C. albicans, particularly macrophage IL-1β responses. Consequently, we further dissected the underlying mechanisms of how albumin exacerbates IL-1-driven inflammatory responses in macrophages.

Albumin enhances C. albicans-induced NLRP3 inflammasome activation

Given the central role of IL-1β in driving innate and adaptive immune responses^39,40, its processing and release is tightly regulated⁴⁰. In macrophages, IL-1β processing is a two-step process requiring priming and activation of the NLRP3 inflammasome (Fig. 3A)²¹. We hypothesized that albumin may increase C. albicans-induced NLRP3 inflammasome activation in macrophages, thereby driving their hyperinflammatory response (Fig. 2C). Macrophages were primed with lipopolysaccharide (LPS) to increase pro-IL-1β and pro-IL-18 levels, and were subsequently stimulated with C. albicans to activate the inflammasome. Both IL-1β and IL-18 release induced by C. albicans increased in the presence of albumin (Fig. 3B). Replacing the LPS priming step with albumin priming did not significantly drive IL-1β responses to 5 h C. albicans stimulation (Supplementary Fig. S6A), suggesting that albumin rather boosts inflammasome activation than inflammasome priming. Dose response experiments show that albumin enhances C. albicans elicited IL-1β responses from concentrations above 0.5 mg/mL in LPS-primed macrophages and 5 mg/mL in unprimed macrophages (Supplementary Fig. S6B). Co-stimulation of pattern recognition receptor ligands with albumin in both LPS-primed, as well as unprimed macrophages did not induce substantial (exceeding 100 pg/mL) IL-1β responses, excluding that synergy between albumin and pathogen-associated molecular patterns (PAMPs) drives IL-1β responses (Supplementary Fig. S6C). Similarly, enhanced IL-1β release was observed in murine bone-marrow derived macrophages (BMDMs) stimulated with C. albicans in the presence of either human (Fig. 3C and Supplementary Fig. S6D) or murine albumin (Supplementary Fig. S6E), although less potently.

Fig. 3. Albumin boosts C. albicans-induced NLRP3 inflammasome activation in macrophages.

A Overview of NLRP3 inflammasome priming and activation. DAMPs damage-associated molecular patterns, TLR Toll-like receptor, PAMPs pathogen-associated molecular patterns. B IL-1β and IL-18 release by LPS-primed hMDMs (n = 14) in response to C. albicans infection in the presence or absence of albumin (10 mg/mL). C IL-1β release by LPS-primed mBMDMs (n = 10) in response to C. albicans infection in the presence or absence of albumin (10 mg/mL). D IL-1β release by LPS-primed hMDMs in response to C. albicans infection in the presence or absence of albumin (10 mg/mL) and the inflammasome inhibitors KCl (n = 8 donors) or VX-765 (n = 9 donors). E Percentage ASC-positive mBMDMs (n = 3 uninfected, n = 4 infected). F, G IL-1β release of LPS-primed mBMDMs deficient in the NLRP3 receptor (Nlrp3^−/−), the adapter proteins ASC (Pycard^−/−) or the proteolytic enzyme caspase 1 (Casp1^−/−/Casp11^−/−) (n = 4) (F) or deficient in gasdermin D (Gsdmd^−/−) (n = 3) (G). Bars represent the mean + SEM with dots as individual biological replicates: Individual donors (B, D) or mice (C, E–G) from at least three independently conducted experiments. Statistical significance was determined using a two-way RM-ANOVA with Holm–Šídák post-hoc test.

Pro-IL-1β processing can either be caused by canonical NLRP3 inflammasome activation or non-canonical IL-1β processing⁴¹. The NLRP3 inflammasome complex consists of three units: the sensor receptor domain NLRP3, the adapter apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) and the proteolytic enzyme caspase-1²¹ (Fig. 3A). To determine whether albumin promotes activation of the canonical NLRP3 inflammasome, potassium efflux, an important trigger of NLRP3 activation, was inhibited by increasing the extracellular potassium concentration with potassium chloride (KCl). Additionally, the role of caspase-1 was determined using the caspase-1 inhibitor VX-765. The albumin-enhanced IL-1β release by hMDMs and mBMDMs was neutralized by both KCl and VX-765 (Fig. 3D and Supplementary Fig. S6F).

Using BMDMs isolated from mice expressing the fluorescent citrine coupled to the ASC adapter protein, inflammasome assembly could be visualized through aggregation of the ASC adaptor⁴². A higher abundance of “ASC specks” was observed in the presence of albumin compared to the control condition (Fig. 3E). BMDMs from mice deficient for each of the inflammasome components, the sensor Nlrp3 (Nlrp3^−/−), the adapter protein ASC (Pycard^−/−), and the proteases caspase-1 and −11 (Casp1^−/−/Casp11^−/−), failed to induce IL-1β release in response to C. albicans, which is in line with previous observations in BMDCs¹⁵. Given that 10 mg/mL albumin was unable to circumvent this dependency, this highlights that albumin-enhanced IL-1β release crucially relies on a functional NLRP3 inflammasome (Fig. 3F).

Besides pro-IL-1β processing, caspase-1-mediated gasdermin D (GSDMD) cleavage enables the formation of GSDMD pores within the macrophage membrane that facilitate IL-1β release as well as a cell death process called pyroptosis^43,44. When stimulating gasdermin D deficient (Gsdmd^−/−) murine BMDMs with C. albicans, the IL-1β release still increased in the presence of albumin, yet this increase was lower than in BMDMs from wild-type (WT) mice (Fig. 3G). This indicates that the IL-1β release in the presence of albumin and C. albicans is partially, but not fully, GSDMD-dependent. Collectively, these data mechanistically link albumin to increased activation, but not priming, of the NLRP3 inflammasome in response to C. albicans.

Albumin-enhanced IL-1β release is driven by C. albicans

To ensure that the albumin-potentiated NLRP3 inflammasome activation is not just a phenotype of the C. albicans reference strain SC5314, the IL-1β response of macrophages to different C. albicans strains was quantified in the presence of albumin. Albumin consistently enhanced macrophage IL-1β responses to the oral commensal 101⁴⁵, and vaginal isolates C81 and CA3153⁴⁶ (Fig. 4A).

Fig. 4. Albumin-enhanced IL-1β release relies on C. albicans and is independent of candidalysin.

A IL-1β release by LPS-primed hMDMs in the presence or absence of albumin induced by the oral commensal isolate 101 (n = 10) and the vaginal isolates C81 (n = 10), CA3153 (n = 7). B IL-1β release by LPS-primed hMDMs in response to C. albicans, when either only hMDMs (left panel; n = 12) or C. albicans cells (right panel; n = 12) were preincubated with albumin (5 mg/mL) prior to infection. C IL-1β release by LPS-primed hMDMs (n = 12) induced by C. albicans WT and a candidalysin-deficient ece1ΔΔ mutant. D IL-1β release by LPS-primed hMDMs (n = 8) induced by synthetic candidalysin. E IL-1β release by LPS-primed hMDMs (n = 10) induced by C. albicans WT or a yeast-locked efg1ΔΔcph1ΔΔ mutant. Bars represent the mean + SEM with dots as individual biological replicates of individual donors from at least three independently performed experiments. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed rank test (A, B) or a two-way RM-ANOVA with Holm-Šídák post-hoc test (C–E).

By exclusively exposing either the macrophages or the fungus to albumin before infection, we found that albumin does not change the capacity of macrophages to activate the inflammasome (Fig. 4B, left panel), but instead potentiated C. albicans to induce stronger inflammasome activation (Fig. 4B, right panel).

To unravel how albumin causes C. albicans to drive stronger inflammasome activation, the role of selected virulence factors known to activate the NLRP3 inflammasome was investigated⁴⁷. Candidalysin, a pore-forming toxin of C. albicans, induces inflammasome activation in macrophages¹⁵. However, a strain lacking the candidalysin-encoding gene ECE1 (ece1ΔΔ) still increased IL-1β release in the presence of albumin (Fig. 4C). In line with previous observations that albumin can neutralize the cytotoxic potential of candidalysin⁴⁸, IL-1β release in LPS-primed macrophages induced by extracellular synthetic candidalysin was neutralized by the presence of albumin (Fig. 4D),

A yeast-locked C. albicans efg1ΔΔcph1ΔΔ mutant poorly induced IL-1β release by hMDMs, whereas a trend towards increased IL-1β release was observed with albumin (Fig. 4E). Consequently, in line with previous findings^15,33,49, the magnitude of macrophage IL-1β responses to C. albicans depends on gasdermin D, C. albicans hyphae and the hyphae-associated toxin candidalysin. Nonetheless, neither filamentation nor candidalysin were identified as the essential drivers of the albumin-enhanced IL-1β responses.

Albumin-enhanced IL-1β release relies on C. albicans secreted aspartic proteinase 1

C. albicans can express a family of secreted aspartic proteinases (Saps)¹³, some of which have been described to mediate inflammasome activation^14,50,51. Individual C. albicans deletion mutants of Saps 1–6 were assessed. Specifically, a C. albicans sap1ΔΔ mutant failed to increase IL-1β release in the presence of albumin (Fig. 5A). This observation was reproduced using another independently generated C. albicans sap1ΔΔ mutant and was restored in a sap1ΔΔ + SAP1 complemented strain (Supplementary Fig. S7A).

Fig. 5. C. albicans Sap1 drives albumin-enhanced IL-1β responses.

A IL-1β release by LPS-primed hMDMs (n = 9) induced by C. albicans WT or Sap1-6 deletion mutants in the presence or absence of albumin (10 mg/mL). B IL-1β release by LPS-primed hMDMs (n = 9) induced by recombinant Saps. C, D IL-1β release by LPS-primed hMDMs induced by rSap1 (n = 9; C) or C. albicans (n = 9; D) in the presence or absence of albumin (10 mg/mL) and pepstatin A (PepA). E IL-1β release by LPS-primed hMDMs (n = 9) induced by rSap1, synthetic candidalysin or the two combined. F Normalized SAP1 expression of C. albicans incubated in medium alone (n = 6), on VECs (n = 6) or with hMDMs (n = 6). Expression data is shown as log₂ fold change (FC) of albumin (5 mg/mL) compared to the control without albumin. G Schematic showing albumin potentiated NLRP3 inflammasome activation by C. albicans Sap1. Data are from at least three independently conducted experiments. H IL-1β, neutrophils, and CFU (only week 1) measured in vaginal lavages of estrogen-treated mice naïve (n = 8) or infected with C. albicans WT (n = 8) or sap1ΔΔ (n = 7) at 1 week or 2 weeks post-infection. I Representative microscopy of cytospins of vaginal lavages of naïve, C. albicans WT, or sap1ΔΔ infected mice. Scale bar represents 50 µm. J two-tailed Pearson correlations of albumin levels with IL-1β concentrations in vaginal lavages of C. albicans WT (n = 8) or sap1ΔΔ (n = 7) infected mice. Bars represent the mean + SEM with dots as individual biological replicates. Statistical significance was determined using a two-way RM-ANOVA with Holm–Šídák post-hoc test (A, D, E), one-way RM-ANOVA with Holm–Šídák post-hoc test (B, C), one-way ANOVA with Holm–Šídák post-hoc test (H) or a two-tailed Wilcoxon matched-pairs signed rank test (F).

To verify Sap1 as a NLRP3 inflammasome inducer in macrophages, hMDMs were stimulated with the recombinant Sap (rSap) 1 protein⁵². In contrast to rSap2 and rSap6, rSap1 induced significant IL-1β release in LPS-primed hMDMs (Fig. 5B). Consistent with a previous study⁵², NLRP3 inflammasome activation was independent of the proteolytic activity of rSap1, as pepstatin A, a well-characterized Sap inhibitor, did not diminish IL-1β release (Fig. 5C). In line with this, the albumin-enhanced IL-1β response to C. albicans was still evident in presence of pepstatin A (Fig. 5D), suggesting that overall Sap proteolytic activity is not essentially required for the enhanced IL-1β release.

As candidalysin presence seemed to regulate the magnitude of the elevated IL-1β responses to C. albicans in the presence of albumin (Fig. 4C), we investigated how candidalysin and Sap1 act together on NLRP3 inflammasome activation. The combination of rSap1 with synthetic candidalysin synergistically induced IL-1β release (Fig. 5E).

As a short exposure of C. albicans to albumin prior to the macrophage encounter was sufficient to enhance its ability to induce IL-1β responses (Fig. 4B), we hypothesized that albumin may induce SAP1 mRNA expression. A tendency for increased SAP1 expression in the presence of albumin was found after 3 h when C. albicans was grown alone or while infecting vulvovaginal epithelial cells (Fig. 5F and Supplementary Fig. S7B). During the interaction with hMDMs, SAP1 expression significantly increased in the presence of albumin (Fig. 5F and Supplementary Fig. S7B). Collectively, these results suggest that C. albicans increases SAP1 expression in response to albumin, which activates the NLRP3 inflammasome in macrophages and promotes IL-1β responses (Fig. 5G).

We validated the association of Sap1 with C. albicans-induced IL-1β responses by infecting mice with WT and sap1ΔΔ C. albicans. One week after infection, we observed that sap1ΔΔ infected mice failed to significantly induce IL-1β responses, as well as neutrophil recruitment to the vaginal lumen, despite a similar fungal burden (Fig. 5H, I). Interestingly, two-weeks post infection, IL-1β responses as well as neutrophils in vaginal lavages were similar to WT (Fig. 5H, I). Correlating IL-1β with albumin levels from the mice, we see significant correlations in mice infected with WT but not sap1ΔΔ C. albicans (Fig. 5J). These data confirm a central role for Sap1 in albumin-driven IL-1β responses early during VVC in vivo.

Albumin promotes fungal fitness during confrontation with macrophages

As albumin acted on the fungus to elicit an elevated SAP1 expression specifically in interaction with macrophages, we systematically assessed how albumin altered C. albicans-macrophage interactions.

C. albicans (both WT and a sap1ΔΔ mutant) showed significantly improved survival rates when encountering macrophages in the presence of albumin (Fig. 6A and Supplementary Fig. S8A). In contrast, a yeast-locked C. albicans efg1ΔΔcph1ΔΔ mutant did not benefit from the presence of albumin (Fig. 6A). We went through different stages of the interaction to pinpoint how albumin supports fungal survival within macrophages. Lower numbers of C. albicans cells were phagocytosed in the presence of albumin after 30 min, irrespective of the strain used (Fig. 6B). Nevertheless, this difference was restored 2.5 h after infection with the C. albicans WT and sap1ΔΔ mutant, suggesting that albumin delays phagocytosis at early stages (Supplementary Fig. S8B). Interestingly, albumin promoted the phagocytosis of the efg1ΔΔcph1ΔΔ mutant 2.5 h after infection (Supplementary Fig. S8B). Importantly, blockade of phagocytosis using cytochalasin D neutralized the capacity of albumin to enhance IL-1β responses to C. albicans (Supplementary Fig. S8C).

Fig. 6. Albumin promotes fungal fitness within macrophages.

A Fungal survival 3 h after infection of hMDMs (n = 8), displayed as percentage CFU of inoculum. B C. albicans phagocytosis by hMDMs 30 min after infection (n = 12). C, D Percentage and length of C. albicans hyphae 2.5 h after macrophage infection WT (C; n = 10 donors) or sap1ΔΔ (D; n = 10). Representative images with scale bars indicate 20 μm. E C. albicans SC5314 (n = 5) and sap1ΔΔ (n = 4) growth under oxidative stress. F Normalized expression of C. albicans oxidative stress resistance genes during the interaction with hMDMs after 30 min (n = 10) and 3 h (n = 7). Expression data is shown as log₂ fold change (FC) of albumin (5 mg/mL) compared to the control without albumin. G Representative histograms (normalized to mode) of p-cat1-GFP mean fluorescence intensity. H IL-1β release by hMDMs (n = 6) induced by C. albicans WT and cat1ΔΔ mutant. I Schematic showing albumin-promoted survival of C. albicans within macrophages, allowing the expression of Sap1. Bars represent the mean + SEM with dots as individual biological replicates. Data are from at least three independently conducted experiments. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed rank test (C, D, F) or by two-way RM-ANOVA with Holm–Šídák post-hoc test (A, B, E, H).

Quantification of the fungal germination rate and hyphae length within macrophages suggested that albumin enables stronger filamentous growth (Fig. 6C, D). Intriguingly, in the absence of macrophages, C. albicans forms shorter hyphae in response to albumin (Supplementary Fig. S8D)⁵³. This suggests that albumin specifically promotes filamentation within macrophages, hinting at an interference with the macrophage capacity to control C. albicans filamentation within the phagosome.

Albumin is known for its antioxidative properties⁵⁴, which could impair the oxidative environment of the phagolysosome, a crucial process in killing C. albicans. Accordingly, the growth-inhibiting effect of H₂O₂, indicated by a delayed time to mid log-growth, was neutralized by albumin in both WT and sap1ΔΔ (Fig. 6E). Several oxidative stress genes were significantly downregulated 30 min after exposing the fungus to macrophages in the presence of albumin but upregulated after 3 h (Fig. 6F and Supplementary Fig. S8E). In line with this, using a reporter strain in which GFP expression is controlled by the promotor of the C. albicans catalase 1 gene (CAT1)⁵⁵, we showed that the presence of albumin activates CAT1 promotor activity (Fig. 6G and Supplementary Fig. S9A, B). Interestingly, a C. albicans mutant lacking the catalase 1 gene (cat1ΔΔ), a crucial factor for detoxifying hydrogen peroxide, lost its ability to enhance IL-1β release in the presence of albumin (Fig. 6H).

Collectively, these data show that albumin promotes C. albicans survival within macrophages independent of the capacity to express SAP1. This is mediated through supporting its oxidative stress resistance. The improved C. albicans survival in macrophages enables increased SAP1 expression. Sap1 subsequently synergizes with filamentous growth and candidalysin to potently activate the NLRP3 inflammasome (Fig. 6I).

Albumin causes dysfunctional neutrophil activation and fungal clearance

Based on the albumin-enhanced macrophage responses to C. albicans, we hypothesized that this could translate to different downstream neutrophil activation. Primary human neutrophils stimulated with hMDM supernatants showed significantly increased surface expression of the multifunctional integrin CD11b and the degranulation marker CD66b. This indicated an elevated neutrophil activation, when macrophages upstream had been exposed to C. albicans and albumin (Fig. 7A, gating strategy in Supplementary Fig. S10).

Fig. 7. Albumin increases downstream neutrophil activation and compromises fungal clearance.

A Neutrophil activation marker expression after stimulation with supernatants from macrophages challenged with C. albicans in the presence or absence of albumin (neutrophils n = 3 donors were stimulated with supernatants from n = 3 macrophage donors). Overview heatmap of all markers (left panel) showing mean fluorescence intensity (MFI), normalized to unstimulated medium control. Additionally, normalized MFIs of CD66b and CD11b are shown as bar graph with SEM (right panel). B Remaining fungal CFUs following encounter with neutrophils in the presence or absence of albumin (left panel, n = 8, mean + SEM with individual biological replicates as dots). Fungal hyphal growth assessed by live-cell imaging in the presence or absence of neutrophils with or without albumin and supplemented IL-1β (Right panel, n = 5 donors). C Spontaneous (left) and C. albicans-induced (right) formation of neutrophil extracellular traps in the presence of albumin (5 mg/mL) and/or IL-1β (10 ng/mL) quantified by live-cell imaging (n = 5 donors). D Representative images of the live-cell analysis of fungal outgrowth and NETosis in (B, C) at 8 hpi, scale bar represents 100 µm. E Luminol conversion by reactive oxygen species produced by neutrophils in response to C. albicans in the presence or absence of albumin (left panel, n = 6) or in response to PMA with albumin injection after 10 min (right panel, n = 7). A, B Bars represent the mean + SEM with dots as individual biological replicates. B, C, E Curves represent the mean ± SEM of the biological replicates (technical replicates in case of C. albicans alone in (B)). Statistical analysis was done using a two-way RM-ANOVA with Holm-Šídák post-hoc test (A, B right panel, C, E) or two-tailed paired t-test (B left panel).

As neutrophils are dysfunctional during RVVC^6,8, we investigated how albumin impacts their antifungal activity. The capacity of neutrophils to restrict C. albicans growth improves by IL-1β stimulation (Fig. 7B). Yet, albumin compromised the ability of neutrophils to control C. albicans outgrowth and even overruled any IL-1β-derived benefits for antifungal activity (Fig. 7B, D). Neutrophil extracellular trap formation (NETosis) is an important mechanism by which neutrophils attack fungal hyphae⁵⁶, yet this is dysregulated during VVC^57,58. While albumin as well as the combination of albumin with IL-1β supported viability of uninfected neutrophils, both conditions promoted NETosis during infection (Fig. 7C, D). The production of reactive oxygen species (ROS) is an important mechanism for fungal clearance by neutrophils. In line with the compromised neutrophil-mediated killing and attenuation of oxidative responses by albumin in macrophages, ROS activity derived from neutrophils was compromised by albumin (Fig. 7E).

Discussion

VVC is characterized by dysfunctional hyperinflammatory responses^6,8. Vaginal albumin concentrations increased upon C. albicans infection and correlated with the presence of proinflammatory cytokines in RVVC patients. We showed that this is not merely a correlation with inflammation, but also that albumin causally exacerbates proinflammatory responses of macrophages towards C. albicans, downstream resulting in increased neutrophil activation. Yet, both neutrophils and macrophages exhibit dysfunctional C. albicans clearance in the presence of albumin. Mechanistically, we found that albumin improves C. albicans fitness against macrophages and neutrophils by boosting fungal resistance to ROS, which depends on albumin-induced oxidative stress resistance as well as albumin’s direct suppression of ROS. In the presence of albumin, C. albicans increased expression of SAP1, which we identified as the factor enhancing NLRP3 inflammasome activation in macrophages in synergy with filamentation and candidalysin. Collectively, we demonstrate a role for albumin in NLRP3-inflammasome-driven hyperinflammation and dysfunctional innate fungal clearance associated with VVC (Fig. 8).

Fig. 8. Graphical summary of how albumin mediates NLRP3-inflammasome-driven hyperinflammation and dysfunctional innate fungal clearance associated with VVC.

1) Increased albumin in the vagina can be caused by vascular permeability in response to acute inflammation or estrogen. 2) Albumin induces adaptations in C. albicans, characterized by increased SAP1 expression and resistance to oxidative stress. 3) This collectively drives enhanced NLRP3 inflammasome activation, leading to 4) IL-1-driven inflammation and 5) neutrophil recruitment and activation. 6) When C. albicans engages with neutrophils in an albumin-containing environment, the neutrophils exhibit increased NETosis yet compromised oxidative responses, resulting in impaired fungal clearance. The histology images on the right show Grocott staining of the C. albicans-infected vaginal lumen and vaginal mucosa (top two images, respectively), as well as hematoxylin and eosin (H&E) staining of the Candida albicans-infected vaginal mucosa in mice. This illustrates that within the murine VVC model neutrophils engage C. albicans in both the vaginal lumen and within the vaginal mucosa.

Macrophages seem an overlooked cell type in VVC, but are abundant in the vaginal mucosa^17–20. We established macrophages as plausible source of the VVC-immunopathogenesis driving cytokine IL-1β, as both epithelial cells of that niche and recruited neutrophils barely released any IL-1β compared to macrophages, when encountering C. albicans. Some epithelial cytokine responses were even reduced by the presence of albumin, which may relate to its capacity of neutralizing candidalysin, the efficient driver of epithelial cytokine responses^48,59. In macrophages, cytokine release, particularly IL-1β, was enhanced by the presence of albumin. With evidence on several layers, we show that albumin increases C. albicans-induced NLRP3 inflammasome activation. Albumin compromised the capacity of macrophages to control intracellular fungal survival. In line with the ability of albumin to neutralize ROS⁵⁴, we observed that albumin protects C. albicans against oxidative stress. In addition, we observed that albumin increases the expression of C. albicans oxidative stress resistance genes. Besides its effect on oxidative stress, albumin promotes global metabolic and transcriptional adaptations in C. albicans and can help to overcome iron limitation^32,60. Therefore, albumin may additionally serve as an iron source, and help overcome nutritional immunity inside the phagosome. Likely, a combination of these effects enables C. albicans to form longer filaments and to survive within macrophages.

We pinpointed C. albicans secreted aspartic protease 1 (Sap1) as the driver of elevated NLRP3 inflammasome activation and showed albumin-increased expression of SAP1, specifically during the encounter with macrophages. Independent of macrophages, a trend towards increased SAP1 expression was observed. Given that oxidative-stress was identified as an inducer of SAP1 expression⁶¹, it may be that albumin and low-grade oxidative stress synergize to induce significant SAP1 expression in macrophages. While SAP1 expression has been linked to the yeast morphology, it was increased after 3 h of infection where the early hyphal morphology is predominating. Interestingly, strong upregulation of SAP1 expression has also been observed in C. albicans hyphae interacting with neutrophil extracellular traps⁶². This suggests that gene expression of C. albicans virulence factors cannot by default be dogmatically linked to a specific morphology. In line with this, the hyphae-associated gene ECE1 was found to be expressed by pseudohyphae from VVC patients that simultaneously express yeast-associated SAP2⁶³. We hypothesize that, following exposure to albumin, C. albicans hyphae may deviate from their morphology-coupled expression pattern and thus may be able to sustain (traditionally yeast associated) SAP1 expression even in the hyphal morphology. Supporting this, we found that albumin moderates hyphal growth⁵³ and induces broad transcriptional and metabolic reprogramming of C. albicans associated with downregulation of typical hyphae-associated genes³². Collectively, when albumin is present, the changes in the hyphal program appear to maintain SAP1 expression during the early hyphal stage.

In line with the strong link between C. albicans filamentation and inflammasome activation³³, a yeast-locked efg1ΔΔcph1ΔΔ mutant induced poor IL-1β responses. Yet still, a trend towards increased IL-1β was observed in presence of albumin. We hypothesize that an interplay between phagosome stress imposed by the filamentous growing C. albicans, its resistance to oxidative stress within the macrophage, the expression of SAP1, and potentially also intracellular candidalysin synergize to drive the increased IL-1β responses. Supporting this, an ece1ΔΔ mutant still showed the phenotype of enhanced IL-1β responses in the presence of albumin, despite overall significantly lower IL-1β responses. Furthermore, stimulation of macrophages with rSap1 and synthetic candidalysin synergistically induced IL-1β responses. Thus, albumin-enhanced macrophage IL-1β responses to C. albicans are multifactorial and linked to (1) improved fungal survival within macrophages, (2) the capacity to engage in hyphal growth, and (3) SAP1 expression.

Linking this to clinical VVC, SAP1 was found to be upregulated in C. albicans isolates from VVC patients but not in isolates from healthy women^64,65. Similarly, strains capable of inducing VVC in a rat model showed SAP1 expression early during infection⁶⁶. The general importance of Saps for C. albicans-instigated immunopathology during VVC has been shown in numerous studies^7,50,64. Sap2, Sap5, and Sap6 are characterized inflammasome inducers^14,22,51, yet also Sap1 was shown to increase IL-1β release in monocytes⁵². Deletion mutants of SAP1 also inflict less damage to reconstituted human vaginal epithelium⁶⁷ coinciding with reduced cytokine responses⁶⁸. Our data complements previously published literature by underscoring the role of C. albicans secreted aspartic proteases in IL-1β responses during VVC.

The NLRP3 inflammasome drives neutrophil-mediated inflammation during (R)VVC^7,22,23. Such IL-1-driven inflammatory responses during C. albicans infections trigger neutrophil recruitment, which normally mediate resistance to infection^62,69. For this reason, the interplay between C. albicans, human serum albumin and macrophages can also play roles beyond vaginal candidiasis, and may have implications for systemic infection as well. However, during (R)VVC neutrophils are believed to aggravate disease, rather than effectively clear the infection^4,6,8. We found that albumin compromises the ability of neutrophils to effectively clear C. albicans. This is associated with the antioxidant effects of albumin neutralizing neutrophil-derived ROS directly, as well as increasing fungal oxidative stress resistance. Importantly, the attenuated capacity to antagonize fungal growth was despite increased NETosis. This fits the observation of a high abundance of neutrophil extracellular traps in vaginal discharges of VVC patients⁵⁷. Mechanistically, neutrophil MPO-derived hypochlorous acid can modify albumin, which has been previously identified as a trigger of NETosis⁷⁰.

While the albumin measurements in the RVVC patients provided a strong rationale to dissect mechanistic links between albumin and inflammation, these cohorts and their sampling approach were not designed to associate albumin with specific phenotypes of RVVC nor determine native albumin concentrations in vaginal fluid, respectively. To completely understand the clinical role of albumin and its link to inflammation severity in the multifactorial disease RVVC, specifically designed patient studies are required. As determination of native protein concentrations in cervicovaginal fluid is challenging⁷¹, such studies should include a specific sampling design to correct for dilution factors⁷². Well-designed studies could address whether albumin levels in RVVC patients remain elevated between the symptomatic episodes, and potentially link albumin with specific Candida species and symptom severity. As albumin can modulate drug pharmacokinetics⁷³, it would also be important to investigate how albumin variation affects the antifungal treatment efficacy of VVC.

Our findings, in line with previous studies, reflect the complexity and multifactorial pathogenesis of VVC. The immunopathogenesis of VVC can hardly be pinpointed to a single factor but is a complex combination of different factors and effects on the host side as well as the fungal side^6,8. By combining observations from RVVC patients and a murine VVC model with in vitro mechanistic dissection, we identified that albumin can serve as a catalyst of IL-1β-mediated inflammation. This may underlie inflammation in VVC patients, yet it cannot be excluded that this mechanism also plays roles beyond VVC during invasive candidiasis.

Methods

Ethics statement for primary human & murine macrophages and mouse in vivo experiments

All methods were conducted in accordance with relevant guidelines and regulations. Experimental protocols for murine in vivo studies were approved by the Ministry of Health (Authorization N. 725/2019-PR) and previously certified by the animal ethics committee “OPBA” from the University of Perugia, Italy.

Human blood was 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-Kommission des Universitätsklinikums Jena, Permission No. 2207–01/08 and 2026-4182-BO-A).

Murine bone marrow was obtained in compliance with the German animal protection law and the collection of bone marrow was approved by the local ethics committees of the University clinics of Bonn and Erlangen, respectively.

Collection of human vaginal fluids

Two RVVC patient cohorts were used. The first cohort (RVVC₁), which was kindly provided by Prof. Luigina Romani, consists of RVVC patients (n = 13) and control volunteers (n = 9)³⁴, who enrolled in a study approved by the University Ethics Committee of Perugia (Prot. 2012-028) and provided written informed consent for study participation. Vaginal fluid collection was performed at the S. Maria Della Misericordia Medical Center (Perugia, Italy). Cervicovaginal samples were obtained from all participants by instilling 3 mL of sterile saline into the posterior vagina, mixing the saline with secretions and withdrawing the solution with a syringe. All vaginal washes were centrifuged at 12,000 × g for 10 min to separate the mucus from the PBS wash solution shortly after collection and washes were immediately frozen at −20 °C³⁴. Recruited RVVC patients had symptomatic infection at the time of sampling. The cohort consisted of Caucasian women who were previously diagnosed with at least three VVC episodes in a year, which were microbiologically validated, and caused by C. albicans. Presence of Lactobacilli, Gardnerella vaginalis, and Streptococcus agalactiae were validated based on culture. Furthermore, the sample was examined microscopically using both a fresh (wet mount) preparation and a Gram-stained preparation.

Controls consisted of age-matched healthy Caucasian women with no gynecological complaints, no history of vaginal candidiasis infection, and who were culture-negative for vaginal pathogens. Exclusion criteria were pregnancy, diabetes mellitus, endocrine or immune deficiency disorders, use of immunosuppressive medications, antibiotics or high estrogen content contraceptives, chemotherapy or prior hysterectomy³⁴.

The second cohort (RVVC₂) consisted of RVVC patients (n = 102) with active infection from a previous prospective multicentric follow-up study performed to evaluate the effectiveness of individualized decreasing fluconazole maintenance therapy in preventing relapses of Candida spp. vulvovaginitis (ReCiDiF trial)³⁵. At the time of sampling presence of Candida spp. was confirmed by microscopy, culture, and/or PCR. Vaginal rinsing with 2 mL of 0.9% sodium chloride was performed by flushing and re-aspirating the fluid through a 0.5-mm wide, 6-cm long needle in the left, central, and right upper vaginal vaults³⁶.

Multivariate analysis of patient cohort

Multivariate structure across clinical, microbiological, and immunological variables was explored using Principal Component Analysis (PCA), implemented via the FactoMineR package in R (version 4.5.0)⁷⁴. Prior to analysis, cytokine concentrations were log-transformed to account for skewed distributions. The severity score was excluded from the PCA biplot input to avoid circularity, as it is arithmetically derived from the symptom score variables already included in the analysis. Missing values were replaced by the median for continuous variables and the mode for ordinal variables prior to analysis. PCA results are presented as a variable biplot, in which arrow direction and length reflect each variable’s loading and contribution to the principal components, and as a patient scores plot in which individual patients are positioned according to their component scores and colored by clinical severity.

Experimental design of murine VVC model

Female sex 8 weeks old C57BL/6J mice were housed with a 14 h light and 10 h dark cycle between 22–24 °C and at 45–70% humidity. Mice were treated subcutaneously with 0.2 mg of β-estradiol 17–valerate (Merck) dissolved in 100 µL of sesame oil (Merck) 24 h before vaginal infection and on the day of vaginal infection. Estrogen administration continued once per week until the end of the experiment to maintain pseudoestrus. Mice were inoculated intravaginally with 20 µL of phosphate-buffered saline (PBS) or PBS containing 2 × 10⁹/mL viable C. albicans SC5314 or C. albicans sap1ΔΔ yeast cells from early–stationary-phase cultures (i.e., 18 h of culture at 37 °C on Sabouraud-dextrose with chloramphenicol). Vaginal lavages were conducted with sterile PBS with repeated aspiration and agitation (total 100 μL). The course of infection was monitored in individual mice by culturing 100 µL of serially diluted (1:10) vaginal lavages on Sabouraud-dextrose agar with chloramphenicol. CFUs were enumerated after incubation at 37 °C for 24 h and expressed as log₁₀ CFU/mL of lavage fluid.

Cytospins of vaginal lavage fluid were made by diluting 20 µL collected vaginal wash into 180 µL of PBS. The solutions were spun onto microscopy slides through a cytospin chamber at 700 × g for 7 min and minimum rotor acceleration and deceleration. Produced slides were stained with May-Grunwald/Giemsa staining (Sigma-Aldrich). Slides were imaged at 60× magnification on an EVOS® FL Auto Imaging System (Thermo-Fisher Scientific, software Rev 2.0.2094.0).

Tissue homogenates were made by putting the mouse vaginal tissue in tubes containing 1 mL PBS, subsequently the tissue was homogenized using an IKA ULTRATURRAX tissue homogenizer. Tissue homogenate (100 µL) was plated on Sabouraud-dextrose agar with chloramphenicol. CFUs were enumerated after incubation at 37 °C for 24 h and expressed as log₁₀ CFU/100 µL of homogenate. Homogenates were centrifuged for 10 min at 700 × g and the supernatants were used for albumin and IL-1β quantification.

Murine histology

The reproductive tract was collected and fixed in 10% neutral buffered formalin, trimmed in transverse sections to include portions of the vagina, cervix, and uterus and embedded in paraffin. Histologic slides were prepared using routine methods and 4 μm-thick sections were stained with hematoxylin and eosin or Grocott’s methenamine silver stain for enhanced fungal visualization. The stained slides were observed on a BX53 system microscope (Olympus; cellSense V4.5). Alternatively, immunofluorescence to visualize neutrophil tissue influx and fungal presence was performed. For this, 5 µm-thick sections were cut onto adhesive slides (Superfrost Plus Adhesion, Epredia) and deparaffinated by submersion in a row of xylene (2 × 10 min), 100% ethanol (2 × 1 min), 95% ethanol (1 min), 75% ethanol (1 min), tap and distilled water (1 min each). Samples were encircled with an ImmEdge hydrophobic barrier PAP pen (Vector Laboratories) and after brief drying, slides were washed with PBS including 0.3% Triton-X100. They were blocked with 5% goat serum in PBS with 0.3% Triton-X100 for 90 min and, without prior antigen-retrieval, overnight stained with 80 µL primary antibody mix including PBS, 1% BSA, 0.3% Triton-X100, 10 µg/mL anti-Candida (polyclonal rabbit, BP1006, Acris Antibodies/OriGene Technologies) and 5 µg/mL AlexaFluor-647-coupled anti-Ly6G (clone 1A8, #127610, Biolegend) at 4 °C inside a humid chamber. Samples were washed thrice and secondary stained with 5 µg/mL goat-anti-rabbit IgG-AlexaFluor 488 (A-11008, Invitrogen, Thermo Fisher Scientific) and 5 µg/mL DAPI for 90 min at RT in the humid chamber. Samples were washed thrice and embedded in Fluoroshield (Invitrogen). Images were acquired on an Axio Observer Z1 (Carl Zeiss Microscopy) with Zen software (Carl Zeiss Microscopy v3.12.107.02000).

Cytokine and albumin measurements

Human albumin levels were determined using human albumin ELISA kits (Abcam ab108788 for RVVC₁ or ab108787 for RVVC₂), according to manufacturer instructions.

In samples of the RVVC₁ cohort and the healthy control group, levels of IFN-α, IFN-β, IL-1α (IL-1F1), IL-1β (IL-1F2), IL-1Ra (IL-1F3), IL-6, IL-8 (CXCL8), IL-17A, GM-CSF and TNF were quantified using a human discovery Luminex© assay (R&D Systems) and were evaluated on a Magpix instrument (Thermo) with xPONENT software (v4.2.1324.0). For the RVVC₂ cohort and in vitro experiments with human macrophages or vaginal epithelial cells, cytokine levels (IL-1α, IL-1β, IL-1Ra, IL-6, IL-8, IL-17, and TNF) in vaginal lavages were quantified by DuoSet ELISA (R&D Systems)³⁶.

Murine albumin levels were determined using a murine albumin ELISA kit (Abcam). Vaginal levels of murine IL-1β (IL-1F2) and MPO were quantified using DuoSet ELISAs (R&D Systems). Absorbtion of developed ELISA plates was read in a microplate reader (Tecan Infinite M-Plex; i-control software, v2.0.10.0).

Strains and culture conditions

C. albicans strain SC5314 was used as WT strain if not indicated otherwise. BWP17 + CIP30 was used as the reference strain for experiments with the candidalysin-deficient strain (ece1ΔΔ)¹⁵. Other C. albicans strains used in this study included the oral commensal isolate 101 (kindly provided by Salomé LeibundGut-Landman, University of Zurich)⁴⁵, the vaginal isolate C81 (kindly provided by Julian Naglik, King’s College London), the vaginal isolate CA3153⁴⁶ (kindly provided by Patrick van Dijck, KU Leuven), a yeast-locked C. albicans mutant (efg1ΔΔcph1ΔΔ)⁷⁵ and C. albicans SAP1-SAP6 deletion strains (sap1ΔΔ, sap2ΔΔ, sap3ΔΔ, sap4ΔΔ, sap5ΔΔ, sap6ΔΔ)⁷⁶. CAI4 + Clp10⁷⁷ was used as reference strain for the validation experiment with an independently generated SAP1 deletion strain (sap1ΔΔ)⁷⁸ and SAP1 complemented strain (sap1ΔΔ + SAP1)⁷⁸, as well as for the CAT1 deficient strain (cat1ΔΔ)⁷⁹. Further, C. albicans GFP reporter strains to quantify the promotor activity of ACT1 (pACT1-GFP)⁷⁷, CAT1 (pCAT1-GFP)⁸⁰, and SOD5 (pSOD5-GFP)⁷⁷ were used. Strains were streaked on Yeast Peptone Dextrose (YPD) agar plates and maintained for two weeks. Single colonies were inoculated in liquid YPD medium and grown overnight, shaking at 30 °C and 180 rpm. Fungal cells were washed three times with phosphate-buffered saline (PBS, pH 7.4) and adjusted to the required cell concentration.

Albumin solutions

Human albumin was obtained from Sanquin Plasma Products B.V. (Albuman). If not indicated differently, human albumin was used for experiments. Murine albumin (Sigma) was used for selected experiments. Albumin was used at a final concentration of 10 mg/mL unless indicated otherwise in the description of the respective assay and figure legend. Other concentrations included 5, 1, and 0.1 mg/mL or a dose response ranging from 0.05 to 10 mg/mL.

In vitro vaginal epithelial infection model

To mimic vulvovaginal mucosal infection, A-431 cells (VECs, Deutsche Sammlung von Mikroorganismen und Zellkulturen DSMZ no. ACC 91) or SW954 cells (ATCC HTB-117) were used. These cell lines have been authenticated via commercial STR profiling (Eurofins Genomic) and checked for mycoplasma contaminations using a PCR mycoplasma test kit (PromoKine). VECs were maintained in RPMI-1640 medium containing L-Glutamine (Thermo Fischer Scientific) and supplemented with 10% heat-inactivated fetal bovine serum (FBS; Bio & SELL) at 37 °C and 5% CO₂. For infection experiments, 1 × 10⁵ VECs per mL were seeded in 96-well plates (200 µL/well) or 6-well plates (3 mL/well) and cultured for 2 days (37 °C, 5% CO₂). At the day of infection, VECs were infected with C. albicans strains at a multiplicity of infection (MOI) of 1, unless stated otherwise, in RPMI-1640 without FBS and with or without 5 mg/mL albumin. Infected VECs were incubated for 24 h (37 °C, 5% CO₂). Afterwards, samples were centrifuged at 200 × g for 10 min, supernatants were collected and stored at −20 °C until cytokine quantification.

Human macrophage differentiation

Human monocytes were isolated from venous blood samples of both male and female sex donors and differentiated to macrophages. In brief, human peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats using density centrifugation with lymphocyte separation medium (1.077 g/mL; Capricorn Scientific). After washing the PBMCs twice in PBS, magnetic automated cell sorting (autoMACS, MiltenyBiotec) was used for monocyte isolation using anti-CD14-labeled beads. hMDMs were differentiated by seeding 1 × 10⁷ monocytes in 175 cm² cell culture flask in RPMI-1640 medium containing L-Glutamine and supplemented with 10% heat-inactivated FBS and 50 ng/mL recombinant human M-CSF (ImmunoTools) for 7 days at 37 °C and 5% CO₂ with a medium refreshment after 5 days. Macrophages were detached by adding PBS containing 10 mM EDTA, incubating the cells for 30 min (37 °C, 5% CO₂) and subsequent cell scraping. If not stated differently, hMDMs were seeded at 4 × 10⁴ cells/well in 96-well plates. For the collection of RNA, 6 × 10⁵ cells/well were seeded in 6-well plates. Macrophages were rested over night before infection.

Murine macrophage cultures

Murine BMDMs were generated from bone marrow cells flushed from femurs and tibiae of SPF C57BL/6J WT, Nlrp3^−/−81, Pycard^-/-82, Casp1^−/−/ Casp11^−/−83,84, or Gsdmd^−/−43 deficient or R26-CAG-ASC-citrine reporter mice⁴² (JAX stock #030744). Isolated cells from one leg were seeded into a 175 cm² cell culture flask. Differentiation was done for 7 days at 37 °C and 5% CO₂ in RPMI-1640 medium containing 10% FBS, 40 ng/mL recombinant murine M-CSF (ImmunoTools) and 1 × Penicillin/Streptomycin (100 U/mL; Gibco) with a medium exchange on day 5. Detachment and seeding of murine BMDMs was done as described for human MDMs.

Macrophage stimulations

For cytokine assays, human MDMs were stimulated with C. albicans at MOI 1. Albumin was added at concentration of 10 mg/mL. Samples were incubated at 37 °C and 5% CO₂. After 24 h, samples were centrifuged at 200 × g for 10 min, supernatants were collected and stored at −20 °C until cytokine quantification.

Inflammasome activation was assessed with human MDMs or murine BMDMs. Macrophages were primed with 50 ng/mL lipopolysaccharide (LPS; from Escherichia coli, Sigma) 2 h before infection if not mentioned differently. For inhibitor experiments, Anakinra (recombinant human IL-1Ra, 10 μg/mL, Kineret), potassium chloride (KCl; 25 mM; Merck), the caspase-1 inhibitor VX-765 (50 μg/mL; Invivogen), cytochalasin D (10 μM; Sigma) were added 1 h prior to infection. Subsequently, 10 mg/mL or in selected experiments, 0.05, 0.1, 0.5, 1, 5, or 10 mg/mL albumin was added and macrophages were infected with C. albicans at MOI 10. In selected experiments synthetic candidalysin (20 μM; Peptide Synthetics) was used to elicit inflammasome activation. For preincubation experiments, macrophages or C. albicans were exposed to 5 mg/mL human albumin 1 h prior to infection. For C. albicans preincubation, fungal cells were harvested by centrifugation at 3000 × g washed twice with PBS and adjusted to the concentration for infection. For macrophage preincubation, supernatant containing albumin was removed and cells were washed once with RPMI-1640 before infection. During indicated experiments, proteinase activity was inhibited using the proteinase inhibitor pepstatin A (Sigma; 15 μM). Infected macrophages were incubated for 5 h (37 °C, 5% CO₂) before harvesting the supernatants as described above. For assays with rSaps, Sap1, Sap2, or Sap6 were added at a concentration of 20 μg/mL after LPS priming and assay time was extended to 24 h. For combination experiments of Sap1 with candidalysin, synthetic candidalysin (20 μM; Peptide Synthetics) was used.

Inflammasome activation was assessed with human MDMs stimulated with different PAMPs (10 μg/mL β-glucan, 10 μg/mL chitin, 10 μg/mL mannan, 10 μg/mL zymosan, 100 ng/mL flagellin, 100 ng/mL Pam3Cys and 100 ng/mL poly I:C) in the presence and absence of 10 mg/mL albumin. Macrophages were either primed with or without 50 ng/mL LPS 2 h prior to stimulation. LPS-primed macrophages were incubated for 5 h (37 °C, 5% CO₂) while non-LPS-primed macrophages were incubated for 24 h (37 °C, 5% CO₂) before harvesting the supernatants as described above. Supernatants were then stored at −20 °C until cytokine quantification.

Inflammasome activation was assessed with human MDMs primed with different stimuli (50 ng/mL LPS, 5 mg/mL albumin and 10 mg/mL albumin) 2 h before infection. Macrophages were subsequently infected with C. albicans at MOI 10 for 5 h (37 °C, 5% CO₂) before harvesting the supernatants as described above. Supernatants were then stored at −20 °C until cytokine quantification.

Live-cell microscopy of macrophage viability

hMDMs were seeded in 96-well plates at 4 × 10⁴ cells/well and were incubated in RPMI-1640 medium supplemented with 0.5 µM SYTOX Green (Invitrogen) and were allowed to settle for 15 min at 37 °C with 5% CO₂. Macrophages were cultured in the presence or absence of human albumin (10 mg/mL) Wells were imaged in an Incucyte SX5 Live-Cell Analysis System (Sartorius AG). Phase contrast, green (excitation, 453–485 nm; emission, 494–533 nm) images were taken every hour for 24 h at 20× magnification. The Neurotrack tool’s nuclei parameter was used to identify SYTOX Green-positive macrophages (Incucyte software 2024B; Sartorius). The analysis definition can be shared upon request.

Macrophage metabolic measurements

Albumin influence on macrophage central metabolism was assessed by Seahorse (Agilent). Primary hMDMs (4 × 10⁴ per well) were seeded as described above, but in 8-well Seahorse cell culture plates the day before the experiment. The following day, cells were washed with and their medium exchanged to Seahorse RPMI-1640 pH 7.4 (Agilent), containing 11 mM d-Glucose, 2 mM l-Glutamine, and 1 mM Sodium pyruvate (“seahorse medium”). Cells were then exposed to 10 mg/mL human albumin or left untreated for 1 h at 37 °C without CO₂. Afterwards, cells were washed again and medium exchanged to fresh Seahorse medium, before the cells were subjected to a MitoStress assay procedure in a Seahorse Mini instrument (Agilent) with XFe assay (v3.0.0.41). Throughout sequential injection of Oligomycin (1 µM final), FCCP (10 µM final) and Antimycin A (2.5 µM final) with Rotenone (1.25 µM final) over approximately 90 min, oxygen consumption rate and extracellular acidification rate were continuously measured.

Macrophage surface markers

Albumin influence on macrophage surface markers was assessed by flow cytometry. Macrophages were seeded in a 6-well plate at a concentration of 0.5 × 10⁶ cells/well and incubated overnight in RPMI-1640 supplemented with 10% FCS and 50 ng/mL M-CSF, at 37 °C and 5% CO₂, followed by stimulation in RPMI-1640 with 10 mg/mL of Albumin in a final volume of 2.5 mL for 60 min. Macrophages were detached by using 10 mM EDTA in PBS for 15 min at 37 °C, followed by gentle scraping. Macrophages were collected by washing the wells with 2% FCS in PBS (FACS buffer) and centrifuged at 300 × g, 4 °C for 10 min. Cells were incubated with FACS buffer for 15 min at 4 °C, followed by staining of surface markers for 20 min with antibodies against the following targets: CD86-AF488 (IT2.2, Invitrogen), CD273-PE (MIH18, BioLegend), CD14-BV421 (M5E2, BioLegend), CD206-BV786 (19.2, BD Biosciences), CD163-BV711 (GHI/61, BD Biosciences), HLA DR-APC-eFluor 780 (L243, Invitrogen), CD64-Alexa Fluor 700 (10.1, R&D Systems), CD80-PerCP (MEM-233, Invitrogen). Flow cytometry was performed on an LSR Fortessa (Becton Dickinson) with BD FACSDiva Software (v8.0.1), and data were analyzed with FlowJo v10.6.2 software.

Neutrophil isolation

Neutrophils were isolated from freshly drawn venous blood. Blood was diluted 1:1 with sterile PBS, and cells were separated by layering lymphocyte separation medium (1.077 g/mL; Capricorn Scientific) under the blood and centrifugation for 30 min at 700 × g and room temperature. Plasma, PBMCs, and lymphocyte separation medium were removed and hypotonic lysis buffer was added to the layer of erythrocytes and granulocytes. Samples were incubated on ice for 15 min and subsequently centrifuged at 300 × g and 4 °C for 10 min. Supernatant was removed and the lysis step was repeated for 10 min. The pellet was washed twice with ice-cold PBS and centrifuged at 300 × g and 4 °C for 10 min. After washing, cells were resuspended in RPMI-1640 and adjusted to the required assay concentration.

Flow cytometric analysis of neutrophil activation

Human MDMs were stimulated with RPMI-1640 with or without 10 mg/mL albumin and infected with C. albicans infection at a MOI of 1. After 24 h, macrophage supernatants were used to stimulate human neutrophils. Neutrophils were seeded at a density of 2 × 10⁵ cells/well in a 96-well plate. After 2 h of stimulation, neutrophils were washed using cold FACS buffer (PBS containing 2% FBS). Neutrophils were preincubated with Fc-Block Human TruStain FcX (BioLegend) and stained using fluorophore-linked antibodies against CD15-APC/Fire750 (SSEA-1) and the activation markers CD62L-AlexaFluor647 (DREG-56), CD35-FITC (E11), CD66b-PE (G10F5), CXCR2/CD182-PE-Cy7 (5E8), CD16-PerCP-Cy5.5 (3G8), and CD11b-BV421 (ICR44; all from Biolegend). To exclude dead cells, fixable viability dye eFluor506 (Invitrogen, Thermo Fisher Scientific) was included. Neutrophils were stained for 20 min at 8 °C in the dark, cells were washed in FACS buffer and filtered through a 70 μm mesh before measurement on a FACSVerse Cell Analyzer (BD Biosciences) with BD FACSuite Software (v1.0.6). FlowJo v10.6.2 software. was used for analysis, gating strategy is depicted in Supplementary Fig. S9.

Fungal killing by neutrophils

Human neutrophils were seeded in 96-well plates at a concentration of 5 × 10⁴ cells/well and infected with C. albicans MOI 0.5 in the presence or absence of 5 mg/mL albumin. Cells were incubated at 37 °C and 5% CO₂. After 3 h, the content of the wells was scraped, resuspended and transferred to a collection tube. Afterwards, sterile H₂O was added to each well and remaining content of the wells was lysed, resuspended and transferred to the respective collection tube. The procedure was repeated four times. Cell suspensions were diluted in PBS and plated on YPD agar in duplicate. After 2 days of incubation at 30 °C, number of colony-forming-units (CFU) was quantified.

Live-cell microscopy of neutrophil inhibition of fungal filamentation, NETosis and death

Neutrophils were seeded in 96-well plates (2 × 10⁴ cells/well) in RPMI-1640 containing 0.5 µM SYTOX Green (Invitrogen) and were allowed to settle for 15 min at 37 °C with 5% CO₂. Neutrophils were infected by adding 50 µL (MOI:1, 2 × 10⁴ cells/well) mScarlet-expressing C. albicans cells⁸⁵. Neutrophils and C. albicans were cultured in the presence or absence of human albumin (5 mg/mL) and/or IL-1β (10 ng/mL). Wells were imaged in an Incucyte SX5 Live-Cell Analysis System (Sartorius AG). Phase contrast, green (excitation, 453–485 nm; emission, 494–533 nm) and orange (excitation: 546–568 nm; emission: 576–639 nm) images were taken every 20 min for 16 h at 20× magnification. Hyphal length was quantified with the NeuroTrack analysis tool (Incucyte software 2024B; Sartorius)⁸⁶. The tool’s nuclei parameter was used to identify SYTOX Green-positive NETosis events. The analysis definition can be shared upon request. Following incubation, culture supernatants were collected for IL-1β ELISA.

ROS production by neutrophils

Neutrophils were seeded in white 96-well plates at 5 × 10⁴ cells/well, and stimulated by addition of 50 μL C. albicans 1 × 10⁶ cells/mL with or without 5 mg/mL albumin, or 250 ng/mL Phorbol 12-myristate 13-acetate (PMA; Sigma). Luminol-horseradish peroxidase (HRP) solution (200 µM luminol and 16 U HRP) was added (100 μL) and chemiluminescence was measured every 2.5 min for 3 h at 37 °C in a microplate reader (Tecan M-Plex, iControl software v2.0.10.0). For assays with PMA, albumin (5 mg/mL) was injected 10 min after initiation of measurements.

ASC reporter assay

BMDMs from R26-CAG-ASC-citrine mice⁴² were adjusted to a concentration of 2 × 10⁵ macrophages/mL and 300 μL of this cell suspension were seeded in an 8-well microscopy slide (ibidi) and were rested overnight. Macrophages were primed for 2 h with 50 ng/mL LPS before adding C. albicans SC5314 at MOI 2 in the presence (10 mg/mL) or absence of albumin. Samples were incubated (37 °C, 5% CO₂) and fixed after 5 h using 4% formaldehyde (Carl Roth). Nuclei of macrophages were stained using DAPI (Sigma Aldrich, Merck) to determine total amount of macrophages per image. Microscopic images were acquired at a 20 × magnification using an Axio Observer Z1 (Carl Zeiss Microscopy). Number of macrophages and amount of ASC specks were determined using the counting tool in Zen (Carl Zeiss Microscopy, v3.12.107.02000).

RNA isolation

VECs or hMDMs were seeded in 6-well plates and infected with C. albicans MOI 1 in the presence or absence of 5 mg/mL albumin. Samples for RNA isolation were collected 3 h after infection. As control, samples of C. albicans cells alone were taken at the same time point. RNA sample collection was done by removing the culture media and replacing with 500 μL of RNeasy Lysis buffer (Qiagen), containing 1% β-mercaptoethanol (Roth). Cells were detached using a cell scraper, immediately shock-frozen in liquid nitrogen, and stored at −80 °C until further use. Collected samples were defrosted on ice and centrifuged for 10 min (20,000 × g, 4 °C). Fungal RNA was isolated from the pellet using a freezing-thawing method. RNA quantity and integrity was determined using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, NanoDrop 1000 Operating Software, v3.8.1) and Agilent 2100 Bioanalyzer (Agilent Technologies, Agilent 2100 Expert Software, vB.02.08.SI648), respectively.

Reverse transcription-quantitative PCR (RT-qPCR)

Isolated RNA was treated with DNase I (RNase-Free DNase Set, Qiagen) following the manufacturer’s recommendations and subsequently transcribed into cDNA using 0.5 µg Oligo(dT)12-18 Primer, 200 U Superscript™ III Reverse Transcriptase and 40 U RNaseOUT™ Recombinant RNase Inhibitor (Thermo Fischer Scientific). Obtained cDNA was used for qRT-PCR with GoTaq® qPCR Master Mix (Promega) in a CFX96 thermocycler (Bio-Rad). The expression levels were normalized against 18 s rRNA. All primers used are listed in Supplementary Table S1.

Fungal killing by hMDMs

Human MDMs were seeded in 96-well plates at 4 × 10⁴ cells/well and rested overnight as described above. For infection, macrophages were infected with C. albicans MOI 1 in the presence or absence of 10 mg/mL albumin. Samples were incubated at 37 °C and 5% CO₂. After 3 h, the content of the wells was scraped, resuspended and transferred to a collection tube. Afterwards, sterile H₂O was added to each well and remaining content of the wells was lysed, resuspended and transferred to corresponding collection tubes. The procedure was repeated four times. Cell suspensions were diluted in PBS and plated on YPD agar plates in duplicate. After 2 days of incubation at 30 °C, number of colony-forming-units (CFU) was quantified.

Phagocytosis and intraphagosomal hyphal length assay

Human MDMs were adjusted to a 2 × 10⁵ cells/mL and 300 μL of this cell suspension were seeded in an 8-well microscopy slide (ibidi) and rested before the experiments as described above. Cells were infected with C. albicans MOI 1 and in the presence or absence of 10 mg/mL albumin. To synchronize phagocytosis, samples were centrifuged at 200 × g for 2 min before incubation at 37 °C and 5% CO₂. For quantification of phagocytosis, cells were fixed after 30 min using 4% formaldehyde. For analysis of the intraphagosomal hyphae length, samples were fixed after 2.5 h. Extracellular fungal cells were stained using 50 μg/mL Concanavalin A—AlexaFluor 647 (Invitrogen, Thermo Fisher Scientific). Macrophages were washed twice using PBS and permeabilized by exposing the samples for 5 min to 0.5% Triton X (Sigma). Samples were washed with PBS again and intracellularly located C. albicans was visualized using 50 μg/mL calcofluor white (Fluorescent brightener 28, Sigma). Finally, samples were washed using H₂O and stored in PBS. Microscopy was performed using an Axio Observer Z1 (Carl Zeiss Microscopy) at 20 × magnification. Image analysis was done using the counting tool and the spline curve tool in Zen software (Carl Zeiss Microscopy, v3.12.107.02000). All visible yeasts and hyphae per image were measured to prevent biased analysis.

Growth curve of C. albicans in the presence of albumin and hydrogen peroxide

Growth curves of C. albicans SC5314 and sap1ΔΔ were done by seeding 1 × 10⁴ fungal cells/well in 96-well plates in RPMI-1640. Respective conditions included 5 mg/mL albumin or 1 mM H₂O₂ (Roth) were added. Samples were sealed with foil and incubated at 37 °C in a Tecan M-Plex reader with Magellan Pro v 7.5 software. The optical density at 600 nm was measured every 30 min for 24 h and the plate was shaken for 15 s at 140 rpm before each measurement.

Flow cytometry with GFP reporter strains

GFP-reporter strains were preincubated for 1 h in RPMI-1640 with and without 5 mg/mL albumin at 37 °C and 180 rpm. Subsequently, cells were washed once with RPMI-1640 and incubated for 2 h in RPMI-1640, containing 50 mM H₂O₂ at 37 °C and 180 rpm. Afterwards, cells were centrifuged for 5 min at 5000 × g and fixed with 4% formaldehyde for 15 min. Finally, cells were resuspended in 400 µL PBS and measurements were performed with a FACSVerse (BD Bioscience) flow cytometer. Data analysis was performed using FlowJo v.10.

Statistics and reproducibility

GraphPad Prism 10 was used for analyses. All data presented in the manuscript are the results of at least two independently conducted experiments. Normality testing was performed using the Shapiro–Wilk test. Subsequently, parametric and non-parametric statistical tests were conducted as appropriate and are indicated in the figure legends and significances are depicted in the figures. No statistical method was used to predetermine sample size. Data were excluded from analysis when negative controls (unstimulated) showed cytokine responses, hinting at undesired pre-activation of isolated cells. The experiments were not randomized. The Investigators were not blinded to allocation during experiments and outcome assessment.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Conceptualization: A.D., M.P., T.Z., M.S.G. Methodology: O.E., D.M.D.G., T.P., E.L., A.D., S.Westermann, S.Wirtz, C.B., B.H. Investigation: S.A., K.O.C., A.D., G.V., D.M., S.U.J.H., K.S.S., C.F.F., S.S., M.J., N.J., S.Wisgot, O.E., C.B., G.R., V.O., G.G.G.D., M.P., T.Z. Visualization: S.A., K.O.C., A.D., G.V., D.M., S.U.J.H., T.Z., M.S.G. Funding acquisition: B.H., G.G.G.D., M.S.G. Project administration: S.A., A.D., M.S.G. Supervision: B.H., M.P., A.D., T.Z., M.S.G. Writing–original draft: S.A., M.S.G. Writing–review & editing: S.A., A.D., S.U.J.H., D.D.G., G.G.G.D., B.H., M.P., T.Z., M.S.G.

Peer review

Peer review information

Nature Communications thanks Gordon Brown, Rebecca Drummond, and Riina Richardson for their contribution to the peer review of this work. A peer review file is available.

Funding

M.S.G., S.H., and K.S. were supported by the German Research Foundation (Deutsche Forschungsgemeinschaft—DFG) Emmy Noether Program (project no. 434385622/GR 5617/1-1) to M.S.G. K.C., M.S.G., and G.G.G.D. were supported by a Research Foundation Flanders (FWO)-funded SBO project Defining Vaginal Candidiasis Elements of Infection and Remedy DeVEnIR (project number S006424N). A.D. was supported by an Exploration Grant of the Boehringer Ingelheim Foundation (BIS) to M.S.G. and an ESCMID research grant 2025 to A.D. This research was supported by the Free State of Thuringia and co-funded by the European Union—Project-ID 2023 FGI 0004. “A Live broadcast of the interactions between host and fungal pathogens” to M.S.G. and A.D. M.P. was supported by the DFG within the Collaborative Research Centre (CRC)/Transregio (TRR) 124 “Pathogenic fungi and their human host: Networks of Interaction—FungiNet” project C1 (DFG project number 210879364) to M.S.G. C.B. and S.W. were supported by the DFG within the CRC/TRR 241 “IECinIBD” project A03. S.A. was supported by funding from the European Union’s Horizon 2020 research and innovation program under grant agreement 847507 (HDM-FUN) to B.H. B.H. was further supported by the DFG Priority Program 2225 “Exit Strategies of extracellular pathogens”. We acknowledge support by the Jena School for Microbial Communication (JSMC) that is third-party funded by the Thuringian Ministry of Education, Science and Culture (TMBWK) and the Cluster of Excellence Balance of the Microverse that is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany´s Excellence Strategy—EXC 2051—Project-ID 390713860. The funding providers had no role in study design; collection, analysis and interpretation of data; in the writing of the article; nor in the decision to submit the article for publication.

Data availability

All data are available in the main text and the supplementary materials, and source data. Raw data will be made available upon request. Source data are provided with this paper.

Code availability

R code for the multivariate analysis is available on figshare https://doi.org/10.6084/m9.figshare.33016352.

Competing interests

G.G.G.D. is the chairperson of Femicare vzw and has worked as a medical consultant for various industries. None of these organizations or companies were involved in funding, the design, communication, or data analysis of this study. The other authors declare no competing interests and no conflicting financial interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-78761-w.