INTRODUCTION
The yeast
Candida albicans is normally a harmless commensal that colonizes mucosae of the gastrointestinal tract, oral cavity, and vagina (
1–3). Under pre-disposing conditions,
C. albicans can cause mucosal infections that severely impact quality of life (
4). Oropharyngeal candidiasis (OPC) is the predominant opportunistic oral infection in individuals infected with human immunodeficiency virus (HIV) and is indicative of HIV disease (
5). Vulvovaginal candidiasis (VVC) affects 75% of women at least once during their reproductive years, and more than 5% of women are diagnosed with recurrent vulvovaginal candidiasis (RVVC), having four or more infections annually (
6,
7). Alarmingly, this translates to about 138 million women with RVVC per year globally (
8). While VVC is associated with microbial dysbiosis, high estrogen levels, behavioral practices, and uncontrolled diabetes mellitus, an immunocompromised immune status rarely pre-disposes women to VVC (
6).
C. albicans has several virulence factors including adhesins, invasins, hydrolases, and the ability to transition between a yeast and hyphal morphology (
4,
9). However, epithelial inflammatory and repair responses, as well as mucosal damage and necrotic cell death, are predominantly triggered by the peptide toxin candidalysin (
10–15). Prior to secretion, candidalysin is embedded into a polyprotein precursor, Ece1, which consists of a secretion signal peptide, the precursor peptide for candidalysin, and seven other Ece1 peptides. This structure is likely required to prevent autoaggregation owing to the amphipathic and hydrophobic features of the candidalysin peptide (
16). In fact, synthetic candidalysin spontaneously forms aggregates in aqueous solution (
17).
In oral epithelial cells (OECs), candidalysin-mediated activation of epithelial growth factor receptor (EGFR) induces mitogen-activated protein kinase (MAPK) signaling, resulting in c-Fos transcription factor and MAPK phosphatase-1 activation (
18,
19). This results in the release of inflammatory cytokines and activation of potent innate immune responses (
18–20). The epithelial response against
C. albicans is further augmented by the candidalysin-triggered release of alarmins, anti-microbial peptides, and damage-associated molecular patterns that drive immune cell recruitment (
21).
In contrast to OPC, the candidalysin-induced immune response during VVC is not protective (
22,
23). While neutrophils are recruited in large numbers, they do not promote fungal clearance (
24). This dysfunctionality has been attributed to specific host factors in the vaginal environment, including heparan sulfate, anti-
C.
albicans antibodies, and perinuclear anti-neutrophil cytoplasmic antibodies (
25,
26).
VVC can be prevented or treated using probiotics and/or azoles (
6,
22). Nevertheless, VVC is not always cured, and treatment can be complicated by anti-fungal resistance. Women often experience recurring infections even after anti-fungal treatment (
7). Treatment of RVVC requires maintenance-suppressive azole therapy (
6). Therefore, unlike in OPC where candidalysin induces a protective anti-fungal immune response (
27), the neutralization of candidalysin or modulation of downstream inflammatory responses has been suggested as a therapeutic strategy to prevent immunopathology and symptomology during VVC (
10,
28).
Given the crucial role of candidalysin in causing epithelial damage and driving inflammatory responses that underlie VVC pathogenesis, we combined in vitro infection models with in silico modeling to explore nanobody-mediated neutralization as a potential therapeutic strategy to prevent epithelial damage and inflammatory cytokine release.
DISCUSSION
VVC affects millions of women annually, yet treatment options remain limited, and often recurrence is observed. As pathogenesis involves tissue damage and immunopathology that is caused by the
C. albicans toxin candidalysin (
10), we pre-clinically explored nanobody-mediated neutralization of candidalysin as a therapeutic strategy to treat VVC. We observed that a llama-derived anti-candidalysin nanobody dampened epithelial tissue damage caused both by synthetic candidalysin and
C. albicans-secreted candidalysin during infection
in vitro. We showed that neutralization of cytotoxicity was associated with reduced activation of OECs. In VECs, neutralization resulted in the reduced release of proinflammatory cytokines and reduced neutrophil activation and recruitment. The data suggest that targeting candidalysin therapeutically could break the hyperinflammatory loop that drives VVC immunopathology and severity of symptoms.
Antibody-mediated neutralization of a microbial toxin has previously been explored for the vaginal pathogen
Gardnerella vaginalis (
33). Antibodies against its cytolytic toxin vaginolysin successfully reduced damage of host cells. Similar findings were observed for nanobodies generated against the Shiga toxin of
Escherichia coli (
34).
Here we show the potential of neutralizing candidalysin as a therapeutic strategy. Anti-candidalysin nanobodies not only neutralized OEC damage during infection but also prevented epithelial activation and downstream cytokine release in infected OECs.
Neutralizing toxins offers an opportunity to block key virulence factors that typically activate the immune system. Thus, toxins are ideal vaccine targets as in the case of tetanus (
35,
36). Targeting virulence factors in therapeutics represents a major advantage over traditional anti-microbial therapies as they can be applied without impacting the healthy microbial flora and offer a reduced risk of developing anti-fungal resistance (
37,
38). Recently, the
C. albicans-secreted zinc-binding protein Pra1 was linked to immunopathology and it was shown that RVVC in women can be reduced by inhibiting this fungal factor using zinc treatment (
39).
To treat acute VVC infections, topical azoles or oral fluconazole is typically prescribed (
6,
40). Fluconazole is also used to treat recurrent and severe infections either as a single dose or maintenance suppressive therapy, yet infections persist in a number of patients irrespective of fluconazole treatment (
6,
40–42). However, a drawback of azole therapy is that it also negatively impacts fungi such as
Saccharomyces species, which are beneficial for the prevention of VVC (
43). We show that a candidalysin-neutralizing nanobody exhibits similar efficacy as fluconazole in protecting VECs from cytotoxic damage induced by
C. albicans. The local fluconazole concentration during treatment is approximately 4–8 µg/mL; thus, our dosage (4 µg/mL) falls within the range of what is expected
in vivo during infection (
6). We also show that the nanobody and fluconazole function additively to reduce VEC damage. Combining anti-candidalysin nanobodies with an anti-fungal drug is an attractive treatment option, as this will reduce fungal growth and hypha formation while concomitantly reducing host cell damage and immunopathology driven by candidalysin.
To further investigate treatment strategies, we developed an
in silico model to give insight into the interaction between candidalysin and the anti-candidalysin nanobody, which can be used to further explore nanobody application. The nanobody neutralized candidalysin-induced VEC damage in a ratio ranging from 1:2 to 1:5 (nanobody:candidalysin). Based on
in vitro and
in silico data, we observed that although more effective when applied at the onset of infection, anti-candidalysin nanobodies can effectively reduce VEC damage when added post-infection. Therefore, the nanobodies are able to neutralize candidalysin in the invasion pocket during an established infection (
15,
29). Based on the dynamics of synthetic candidalysin, the amount of “effective” candidalysin that is secreted by
C. albicans hyphae and capable of VEC damage was predicted to be 107 µM after 24 h. Considering the predicted neutralization ratio, it is recommended that the nanobody should be applied at a maximum daily dose of approximately 50 µM. Our
in vitro data indicated that the nanobody may exhibit an even increased efficacy
in vivo, since we observed that nanobodies were more effective at neutralizing damage caused by
C. albicans (multiplicity of infection [MOI] 1) compared to the addition of synthetic candidalysin to epithelial cells. This phenomenon might be explained by spontaneous aggregation and clumping of the synthetic toxin in aqueous solution (
17) and slower and more controlled release of lower concentrations of native candidalysin by
C. albicans hyphae.
Furthermore, when treating host cells with fungal toxin in vitro, we observed that lower nanobody concentrations were effective against 70 µM candidalysin on VECs compared to OECs, where only the highest nanobody concentration reduced epithelial cell damage caused by 16 µM candidalysin, indicating differences, depending on the host cell type. Nevertheless, on both OECs and VECs, even the lowest nanobody concentration was effective at reducing C. albicans-induced host cell damage.
VVC is predominantly caused by
C. albicans (
44), the species in which candidalysin was discovered (
13,
45). Willems et al. (
46) postulated that
C. albicans is the main etiological agent of VVC, since this species vigorously forms hyphae, expresses candidalysin, and causes immunopathology compared to non-
C. albicans (NAC) species. Vaginal infections by NAC species are, however, rising and
Candida glabrata is the second biggest etiological agent of VVC (
44). VECs display distinct transcriptional responses to NAC species, whereas the epithelial response to
C. albicans is primarily driven by candidalysin (
12). The only NAC species with known
ECE1 orthologs are
Candida africana,
Candida dubliniensis, and
Candida tropicalis (
46–48). However, although
ECE1 gene sequences and peptide structures are relatively conserved between
C. albicans strains and
Candida species, the expression pattern of
ECE1 and the biological role of Ece1 for NAC species are unknown (
47–49). In
C. albicans strains, the host cell damage potential of candidalysin is determined by a series of properties including fungal morphology,
ECE1 expression, processing, and secretion (
29,
48,
49). The lack of
ECE1, in addition to morphological differences, further explains why NAC species are generally less pathogenic and do not induce immunopathology as robustly as
C. albicans (
46). Therefore, NAC infections are reported to be less severe, although contrasting findings are reported in literature (
44).
Our data show that the nanobodies act directly on candidalysin and prevent the toxin from causing epithelial membrane damage, since we observed less calcium influx and delayed permeabilization of candidalysin-treated OECs and lipid bilayers, respectively, in the presence of nanobodies. This is further supported by microscopy images showing that the anti-candidalysin nanobodies bound candidalysin within the invasion pocket on VECs without affecting hypha formation and
ECE1 expression. Given that blocking downstream effects of candidalysin, such as EGFR signaling, may lead to potentially contradicting disease outcomes (
18,
50), directly neutralizing candidalysin seems a far more promising approach. In addition to neutralizing epithelial tissue damage, the anti-candidalysin nanobodies dampened inflammatory responses that drive VVC symptoms. Notably, cytokine secretion by VECs was reduced when nanobodies were added 3 h after
C. albicans infection. Reduced epithelial damage most likely accounts for lower secretion of the alarmin IL-1α, which leads to reduced IL-8 and GM-CSF release, similar to what has been described for candidalysin-exposed OECs (
51).
In line with this, we observed that neutrophils exposed to supernatants of
C. albicans-infected VECs secreted less IL-8 and showed reduced activation in the presence of nanobodies. CXCR2, an integral receptor for neutrophil migration, showed reduced expression in the presence of
C. albicans infection. This supported the notion that CXCR2 bound increasing amounts of IL-8 secreted during vaginal infection and was then internalized during migration along the chemokine gradient (
52). This effect was, however, not mitigated by the presence of nanobody. Surprisingly, expression of L-selectin (CD62L), a neutrophil marker of adhesion to endothelial cells and migration (
53), was increased in response to supernatants of infected VECs and lower in the presence of anti-candidalysin nanobodies. CD62L is expected to be inversely regulated with CD11b during granulocyte activation
in vitro (
54–56), since CD62L is rapidly shed during activation (
53). It is, therefore, difficult to determine how the increased CD62L surface expression, which was reduced by the nanobodies, reflects neutrophil activation state. Even though the degranulation marker CD66b was unchanged, CD35, which can be found inside secretory vesicles (
57), was increased on the surface by infected VEC supernatants but decreased in the presence of the nanobodies. Overall, these data show that nanobodies can reduce neutrophil activation in response to vaginal epithelial
C. albicans infection. Comparably, we also showed reduced neutrophil migration during
C. albicans infection in the presence of anti-candidalysin nanobodies.
Our data show that anti-candidalysin nanobodies may represent a potential strategy to treat VVC by dampening VEC damage and associated inflammatory responses. This, in turn, leads to reduced neutrophil activation and recruitment and thus reduced immunopathology. Collectively, this could result in less severe symptomatic VVC episodes. Nanobodies may prove useful in combination therapy with fluconazole to mitigate the fungal burden, which is not cleared by neutrophils, in parallel with candidalysin-induced inflammation. To successfully implement this therapeutic strategy, future work should prioritize preparing a nanobody formulation that can be applied
in vivo. Various considerations are needed to develop such a therapy including pharmocokinetic and stability studies (
58). Nevertheless, by combining wet bench techniques with bioinformatic modeling, we were able to provide a pre-clinical proof of concept as basis for anti-candidalysin nanobody therapy. Based on our
in vitro data, we established a mathematical model that can support further development of nanobody-mediated candidalysin neutralization in terms of identifying optimal doses and dosing intervals.
MATERIALS AND METHODS
Culture and maintenance of C. albicans
C. albicans strain SC5314 (
59) and BWP17/CIp30 (isogenic to SC5314) were cultured on 1% yeast extract, 2% peptone, and 2% dextrose (YPD) medium (for solid medium, 1.5% agar was supplemented). For infection, a single colony was inoculated into YPD broth that was incubated overnight (ON) at 30°C with shaking at 180 rpm. Yeast cells were washed 3× with phosphate-buffered saline (PBS, pH 7.4). The cell number was enumerated using a Neubauer chamber and adjusted to the number needed for infection.
Culture of human oral and vaginal epithelial cells
TR146 OECs (ECACC 10032305) and A-431 VECs (DSMZ no. ACC91) were cultured in the presence of 10% heat-inactivated fetal bovine serum (Bio & Sell) in Dulbecco’s modified Eagle medium (DMEM)/F-12 medium (Gibco) and RPMI-1640 medium (Gibco), respectively, according to the supplier’s instructions. Cell lines were authenticated by commercial STR profiling (Eurofins Genomic) and checked for mycoplasma contaminations using a PCR mycoplasma test kit (PromoKine) following the manufacturer’s instructions. For all experiments, unless specified otherwise, cells were seeded at a density of 2 × 104/well in 96-well plates and incubated at 37°C and 5% CO2 for 2 days until confluency.
Candidalysin neutralization assays
Two clones of anti-candidalysin single-domain antibodies, specifically called V
HH, but often referred to as nanobody (which is a registered trademark of Ablynx), CAL1-H1 and CAL1-F1 were generated against synthetic and native candidalysin, respectively, as described in reference (
29). CAL1-H1 was produced in
Escherichia coli BL21 and purified from the periplasmic extracts by immobilized metal affinity chromatography on the hist-tag, while CAL1-F1 was produced in
Saccharomyces cerevisiae and purified by affinity chromatography using protA columns. Experiments were first performed on OECs using the
C. albicans SC5314-derived BWP17/CIP30 wild-type strain at a final concentration of 2 × 10
4 cells/well (MOI 1) or 16 µM candidalysin (Peptide Synthetics) in a 96-well plate with a final volume of 200 µL. Strains or peptide toxin were pre-incubated with serial dilutions (4, 8, and 16 µM) of the nanobodies CAL1-F1 or CAL1-H1 for 1 h at 37°C. All mixtures were prepared in serum-free cell culture media. After pre-incubation, mixtures were added to TR146 cells, previously washed 1× with the respective serum-free media. As a control nanobody that does not bind candidalysin, we included a V
HH nanobody, anti-human epidermal growth factor receptor 2.
For neutralization assays on VECs, some modifications were applied. We used
C. albicans SC5314 and candidalysin at a highly lytic concentration of 70 µM (
48) to test the maximum neutralization potential of the nanobody. Nanobodies were either pre-incubated for 1 h at 37°C and shaking at 180 rpm with candidalysin or
C. albicans, simultaneously added with candidalysin or
C. albicans, or added 3 h after candidalysin treatment or
C. albicans infection.
To determine the effect of anti-candidalysin nanobodies on cytokines released by OECs, epithelial cells were seeded at a density of 2.5 × 105/well in 24-well plates, left to reach confluency until the next day, and serum-starved ON before infection. C. albicans SC5314 (MOI 0.01) was pre-incubated with 4 µM CAL1-F1 for 1 h at 37°C while shaking before the mixture was added to OECs in 24-well plates.
After 24 h incubation at 37°C and 5% CO2, plates were centrifuged for 10 min at 250 × g, and supernatants were collected for cytotoxicity and cytokine measurements.
Host cell damage (cytotoxicity)
Epithelial cell damage was quantified by measuring the activity of the cytoplasmic enzyme LDH in the supernatant using a cytotoxicity detection kit (Roche) according to the manufacturer’s instructions.
Quantification of calcium influx
OECs were seeded and serum starved before experiments. The following day, serum-free media were removed from the cells. A solution containing 2.5 µM Fura-2 AM (Thermo Scientific) and 500 µM probenecid (Sigma) was prepared in a saline solution (140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM glucose, and 10 mM HEPES [pH 7.4]). Cells were incubated with the solution containing Fura-2 AM for 1 h at 37°C and 5% CO2 in the dark. In the meantime, candidalysin (70 µM) and nanobody (4 and 16 µM, respectively) mix were prepared in saline solution in a 96-well plate. After sealing the plate, the plate was placed on a microplate shaker in a 37°C incubator for 1 h. Following the 1 h incubation steps, Fura-2 AM solution was removed from OECs, and the saline solution with or without candidalysin and nanobodies was added before readings were taken on a FlexStation 3 multimode microplate reader. Samples were excited at 340/380 nm, and fluorescence was detected at 520 nm. Readings were taken every minute. Results were expressed as a ratio between 340 and 380 nm. Data were normalized to OEC-only controls.
Quantification of bilayer permeabilization
Current measurements were performed using the Orbit 16 system (Nanion) as described previously (
48). In brief, the horizontal bilayers were formed using 1,2-diphytanoyl-sn-glycero-3-phosphocholine lipids in an electrolyte solution containing 0.1 M KCl and 20 mM HEPES at pH 7.4. Candidalysin peptides dissolved in water were added to bilayers at a final concentration of 10 µM. To monitor the effect of anti-candidalysin nanobodies, nanobody was pre-incubated with candidalysin before adding to the bilayer (molar ratio 1:1). Current changes were monitored at a constant voltage of −50 mV for 25 min using Element Data Recorder software (EDR v.3.8.3). Latencies until the membrane permeabilization were quantified using Clampfit v.10.3 (Molecular Devices).
Lysate preparation and western blotting
For western blot experiments, OECs were seeded at a density of 2.5 × 105/well in 24-well plates, left to reach confluency until the next day, and serum-starved ON before infection. C. albicans SC5314 was adjusted to MOI 10 and incubated together with 4 µM CAL1-F1 nanobody for 1 h at 37°C on a shaker before the mixture was added to OECs and incubated for 2 h at 37°C and 5% CO2. Following infections, tissue culture plates were placed on ice; culture medium was removed; and cells were washed with ice-cold PBS. Cells were lysed with 120 µL of RIPA buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, and 5% glycerol) supplemented with protease and phosphatase inhibitors (1:100 dilution, Sigma-Aldrich). Adherent cells were then scraped, transferred into pre-cooled microfuge tubes, and incubated on ice for 30 min. Lysates were clarified by centrifugation at 13,300 × g at 4°C for 10 min. Protein extract concentration was measured using a bicinchoninic acid assay (Thermo Fisher Scientific) according to the manufacturer’s instructions.
Proteins were resolved by electrophoresis on 20% SDS-PAGE gels. Following electrophoresis, proteins were transferred onto nitrocellulose membranes (Bio-Rad). Membranes were blocked in 1× Tris-buffered saline (TBST, Severn Biotech) containing 0.001% Tween 20 (Acros Organics) and 5% skimmed milk powder (Sainsbury’s). After washing once with TBST, membranes were incubated with primary antibody (Table S2) and gentle agitation ON at 4°C. The following day, membranes were washed three times for 5 min with TBST. Membranes were subsequently incubated with rabbit or mouse secondary antibody (Thermo Fisher Scientific) for 1 h at room temperature (RT) and then washed six times for 5 min with TBST. Finally, the proteins were detected using Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore) and developed with an Odyssey Fc Imaging System (LI-COR). Human α-actin was used as a loading control.
ECE1 expression
To quantify ECE1 mRNA expression, A-431 VECs were seeded in six-well plates at a density of 3 × 105/well. After 2 days, confluent VECs were infected with 3 × 105 C. albicans SC5314 cells (MOI 1) in the presence and absence of 4 µM CAL1-F1 nanobody for 24 h. The supernatant was removed from the wells, and 500 µL of RNeasy Lysis (RLT) buffer (QIAGEN) with 1% β-mercaptoethanol (Roth) was added. The well contents were detached using a cell scraper, frozen in liquid nitrogen, and stored at −80 °C until further use. C. albicans cells used as inoculum served as a 0 h control.
RNA was extracted by thawing the samples on ice and centrifugation for 10 min (20,000 ×
g, 4°C). Fungal RNA was isolated from the pellet using a freezing-thawing method, as described previously (
60). RNA concentrations were measured with a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific), and quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). RNA (500 ng) was then treated with DNase I (Fermentas) following the manufacturer’s instructions and transcribed into complementary DNA (cDNA) using 0.5 µg of Oligo(dT)
12–18 Primer, 200 U of Superscript III Reverse Transcriptase, and 40 U of RNaseOUT Recombinant RNase Inhibitor (Thermo Fischer Scientific). cDNA was diluted and used for qPCR with GoTaq qPCR Master Mix (Promega Corporation) in a CFX96 thermocycler (Bio-Rad Laboratories). Expression levels were normalized to the housekeeping gene
ACT1 (β-actin) and expressed relative to the expression of the target gene
ECE1 in 0 h control (log
2 fold change). Primers that were used are listed in Table S3.
Live-cell imaging of vaginal epithelial cell death
PI (Sigma-Aldrich) was used to stain non-viable VECs and monitor necrotic cell death over 24 h, as described previously (
61). In brief, VECs were seeded in a 96-well plate, washed once with serum-free RPMI medium, and infected with 1 × 10
5 C. albicans (MOI 1) in medium with and without 4 µM CAL1-F1. PI was added to the wells at a final concentration of 4 µg/mL. VECs were imaged in a Zeiss Celldiscoverer 7 for 24 h at 37°C and 5% CO
2. Images were taken every 20 min at ×10 magnification in bright field and fluorescence (excitation: 545 nm, emission: 572 nm). Using the threshold function in Fiji(
62) images from the red fluorescence channel were converted to binary images. Macro batch analysis and the Particle Analyzer tool were used to quantify the number of PI-positive nuclei. The percentage of dead VECs was calculated in relation to the number of maximum dead host cells after 24 h.
Imaging C. albicans hyphae
To determine if the anti-candidalysin nanobodies affect C. albicans hyphae, VECs were seeded in 24-well plates at a density of 1 × 105/well. After 2 days, VECs were washed once with serum-free RPMI and infected with 1 × 105 C. albicans (MOI 1) in the absence and presence of 4 µM CAL1-F1. After 6 h of incubation at 37°C and 5% CO2, VECs were washed once with PBS and fixed in 4% Histofix (Carl Roth). Bright field microscopy images were taken at ×20 magnification using a Zeiss Celldiscoverer v.7.
Immunofluorescence staining for localization of nanobodies
For candidalysin immunofluorescence staining, A-431 VECs were seeded on glass coverslips (Ø 25 mm, VWR) in six-well plates at a density of 3 × 105/well. To localize the anti-candidalysin nanobodies, VECs were treated with nanobodies at the onset of infection or 3 h after infection and stained. A-431 VECs were washed with RPMI-1640 medium and infected with 1.5 × 105 C. albicans cells (MOI 0.5). After 3 h of incubation at 37°C and 5% CO2, the medium was removed, and the samples were fixed in 4% Histofix (Carl Roth). When the anti-candidalysin nanobody was added 3 h after infection, it was left to interact with secreted candidalysin for 15 min at 37°C and 5% CO2 prior to fixation. VECs were washed twice with PBS and incubated with concanavalin A-Alexa Fluor 647 (20 µg/mL in PBS, Invitrogen) in the dark for 30 min at RT. The cells were then washed twice with PBS, permeabilized with 0.01% Triton X-100 (Carl Roth) in PBS for 10 min at RT and washed once more with PBS. Samples were blocked with 0.5% bovine serum albumin (BSA) in PBS for 1 h at RT and washed twice with PBS. After washing with PBS, the secondary antibody goat IgG anti-camelid VHH (nanobody)-Alexa Fluor 488 (1:340 in 0.5% BSA in PBS, Jackson ImmunoResearch) was added and incubated for 1 h at RT. Samples were washed twice with PBS, mounted using SlowFade Diamond Antifade (Invitrogen), and visualized using fluorescence microscopy.
LSM 980 confocal microscope with Airyscan 2 detector (Carl Zeiss) equipped with C Plan-Apochromat ×63/NA 1.40 Oil DIC M27 objective lens was used to acquire high-resolved images of fungal and epithelial cells. The lasers at 488 and 639 nm were selected for fluorescence excitation of Alexa Fluor 488 and Alexa Fluor 647 dyes, respectively. An automated alignment was performed to calibrate the Airyscan detector before proceeding with the acquisition phase. The super resolution mode of the Airyscan was used with a calibration of 0.043 µm/pixel and a z-step of 0.170 µm. Fluorescence was detected with an Airyscan detector and super-resolution mode. The reconstruction was done using the Airyscan data processing included in the ZEN software with the automatic strength. All measurements were performed at RT. The images were saved in .lsm file format and then analyzed using Fiji/ImageJ.
Isolation and stimulation of neutrophils
Primary human neutrophils were isolated as previously described (
63) In brief, peripheral blood mononuclear cells were separated from granulocytes and erythrocytes using density gradient centrifugation over Histopaque-1077 (Sigma-Aldrich) in a 50-mL sterile tube. Neutrophils were isolated from the erythrocyte/granulocyte fraction using hypotonic lysis of erythrocytes in 155 mM NH
4Cl and 10 mM KHCO
3. Afterward, neutrophils were washed twice in PBS, resuspended in RPMI-1640 media, and seeded in a 96-well plate at a density of 5 × 10
4–1 × 10
5/well. Neutrophils were stimulated with supernatants (2× diluted in fresh RPMI) of VECs exposed to
C. albicans with and without nanobody for 24 h as described above. As control, neutrophils were stimulated with CAL1-F1 nanobody (4 µM) alone. After stimulation, neutrophil supernatants were collected, and IL-8, an indicator of neutrophil activation, was measured using human enzyme-linked immunosorbent assays (R&D Systems) according to the manufacturer’s instructions.
Neutrophil activation was also assessed by flow cytometry. Neutrophils were seeded in a round bottom 96 well-plate at a density of 2 × 10
5/well and stimulated with undiluted VEC supernatants and CAL1-F1 nanobody (4 µM) alone as control. Supernatants were removed after 3 h, and cells were washed in flow cytometry buffer (PBS, 2% fetal calf serum), which all consequent steps were performed in. To prevent unspecific staining, neutrophils were pre-incubated with Fc-Block Human TruStain FcX (BioLegend) before adding a mix of fluorophore-linked antibodies against the following activation status indicating surface molecules: CD11b-BV421 (ICR44), CD15-APC-Fire750 (W6D3), CD16-PerCP-Cy5.5 (3G8), CD35-FITC (E11), CD62L-AlexaFluor647 (DREG-56), CD66b-PE (G10F5), and CD182-PE-Cy7 (5E8, all from BioLegend). Activation markers were selected based on previous observations of granulocyte responses to fungal pathogens or associated stimuli (
52,
54,
57). Fixable Viability Dye eFluor506 (Invitrogen) was used to exclude dead cells. Staining was performed for 20 min at 8°C. Afterward, cells were washed in flow cytometry buffer, filtered through a 70 µm mesh, and acquired on a FACSVerse Cell Analyzer flow cytometer (BD Biosciences). Analysis was performed in FlowJo v.10. For the gating strategy, see Fig. S6.
Neutrophil staining and migration
Isolated primary human neutrophils were stained with cytopainter green (Abcam). Briefly, 2 µL of cytopainter green stock solution was added to 1 × 106 neutrophils in RPMI and incubated at RT for 10 min in the dark. Stained neutrophils were washed once using Hank’s Balanced Salt Solution with 20 mM HEPES buffer (final pH 7) while being spun down at 300 × g for 10 min. After resuspension in endothelial cell medium (ECM, Promocell), the neutrophil cell number was determined using a Neubauer counting chamber.
Cryopreserved human umbilical cord vein endothelial cells (HUVECs) that were kindly provided by the lab of Alexander Mosig (University Clinic Jena) were expanded until four passages in 150-cm2 flasks using ECM and frozen in liquid nitrogen at a concentration of 1 × 106/mL in ECM with 9% FBS and 7.5% dimethyl sulfoxide. HUVECs from glycerol stocks were cultured in 150-cm2 flasks for 72 h and harvested. Cells were then seeded at a density of 2 × 104 cells in a transwell insert with a 3 µm pore size and incubated at 37°C with 5% CO2. After 48 h, transwell inserts were added to 24-well plates with confluent VECs that were seeded 2 days prior at a density of 1 × 105 cells/well. Medium was refreshed in the transwell inserts (200 µL), and VECs were infected with 1 × 105 C. albicans SC5314 cells (MOI 1) in the absence and presence of 4 µM anti-candidalysin nanobody (total volume 600 µL). Following 18 h of infection, 200 µL of cytopainter green-stained neutrophils (5 × 105 cells/mL) was added to the transwell inserts. Plates were incubated for 2 h at 37°C and 5% CO2, where after images were taken of the wells using a Zeiss Celldiscoverer 7 (excitation: 493 nm, emission: 517 nm). The number of cytopainter green-positive events was determined using thresholding similar to that described above for PI image analysis. For each condition, neutrophil migration was quantified as a percentage of the total amount of neutrophils.
Cytokine release
IL-1α, IL-6, IL-8, GM-CSF, and G-CSF were quantified in cell culture supernatants from OECs and candidalysin-treated VECs using magnetic microparticles (R&D Systems) with a magnetic Luminex performance assay (Bio-Techne) and a Bio-Plex 200 system (Bio-Rad) according to the manufacturers’ instructions. Data were analyzed using Bioplex Manager v.6.1 software. Supernatants from C. albicans-infected VECs were analyzed for additional cytokines (IL-8, IFN-α, IFN-β, CCL2, CCL3, CCL4, CCL5, CCL20, IL-1α, GM-CSF, G-CSF, IL-17, CXCL1, and CXCL2) using a multiplex human cytokine panel (R&D Systems) and the Luminex MAGPIX (Thermo Fisher Scientific) instrument according to the manufacturers’ instructions. Any other cytokines released were measured with commercially available human enzyme-linked immunosorbent assay kits (R&D Systems) according to the manufacturers’ instructions.
Nanobody efficacy compared to fluconazole
The efficacy of the anti-candidalysin nanobody was compared to that of FLU, an azole frequently used to treat VVC. A-431 VECs were infected with 2 × 10
4 C. albicans cells (MOI 1), and after 9 h, CAL1-F1 nanobody (4 µM) and/or FLU (4 µg/mL) was added. After a total of 24 h of infection at 37°C and 5% CO
2, 96-well plates were spun down for 10 min at 250 ×
g, and supernatants were collected for cytotoxicity and cytokine measurements. To determine the combined effect of FLU (A) and nanobody (B) on epithelial cell damage, the CDI was calculated as AB/(A × B) using LDH data (
64). After subtracting low controls AB, A, and B were expressed as fold change of the control group.
In silico model description
Our in silico model is based on ordinary differential equations consisting of eight entities. All parameters and entities in the model are listed with their respective units in Table S4. Due to the fact that candidalysin needs to form multimeric aggregates to induce host cell lysis, the entity does not represent the concentration of a single monomer but a group of monomers consisting of the number of entities that are needed to form an aggregate. By making this assumption, the model effectively linearizes the process of aggregation or polymerization to simplify complex multimerization events. This linear approximation allows the model to capture the dynamics observed in the data while maintaining a manageable level of complexity.
The rate constant in the model’s first two equations is scaled by a factor of , which ensures that the effective rate of association between the nanobody and individual monomers of candidalysin supports the simplification of multimerization events:
In these equations, parameters
and
represent the formation rate of the candidalysin aggregate and degradation rates of the nanobody, respectively. The function
serves as a source term for candidalysin secretion by invasive filamentous
C. albicans cells
over time. While this function is not needed when modeling experiments with synthetic candidalysin, it is used to capture experimental data for
C. albicans-secreted candidalysin. Parameter
represents the blocking of candidalysin monomers and aggregates by anti-candidalysin nanobodies. The rate
in model
equations 1 and
2 is scaled by a factor of
to ensure that the effective rate of association between the nanobody and
individual monomers of candidalysin supports the model’s simplification with regard to aggregation representation. In addition, we assumed that the nanobody targets the candidalysin monomer and aggregate with the same binding affinity in a 1:1 ratio; i.e., the nanobody has the same binding affinity to the aggregate as to individual monomers.
The transformation of candidalysin into the aggregate through a reaction involving monomeric candidalysin and its depletion through a reaction involving VECs is given by
Here, is depletion of the candidalysin aggregate due to integration into the host cell membrane subsequently causing damage. The variable α is a conversion constant describing the amount of aggregated candidalysin being depleted relative to 1% of dead VECs.
The damage of VECs caused by aggregated candidalysin is given by
while the production of LDH due to damaged VECs and its degradation rate is given by
Here, the parameter β is a conversion constant that describes the amount of LDH released upon host cell death.
To model the system with C. albicans-secreted candidalysin, we added three more equations to include invasion by C. albicans cells over time:
In these equations,
is the transition rate of yeast cells
into non-invasive filamentous cells
and
is the transition rate from
into invasive filamentous cells
. These processes were modeled previously on OECs, and the corresponding parameters were estimated as previously described in detail (
31).
Furthermore, the source term in
eqation 1 was set to
where refers to the secretion rate of candidalysin. We modeled this process as an interaction term between alive VECs and filamentous invasive C. albicans cells. Therefore, only the candidalysin that was able to cause damage was secreted, which implies that is a secretion rate of “effective” candidalysin. Since the parameters , , and are taken from the synthetic candidalysin experiments, the parameter should be more understood as secretion of candidalysin that is equivalent to synthetic candidalysin.
In silico model parameter estimation
Our
in silico model incorporates various parameters (Table S4) that govern the temporal dynamics of the host damage marker LDH. These parameters cannot be directly observed; therefore, their estimation was based on fitting the model to
in vitro experimental data (
Fig. 2A through C; Fig. S2). To capture the dynamics observed in the data, we utilized a bottom-up approach. The model’s prediction discrepancy was evaluated by comparing it with (i) the LDH concentration and (ii) the percentage of dead VECs minimizing the sum of squared errors (SSE). The percentage of dead VECs is not directly present in the data, but as reference, we can define that this value should be at least as high as the LDH in the
in vitro data as a fraction of maximum LDH release if all VECs are dead (estimated by
). This resulted into the following discrepancy measure:
Our modeling approach includes four model parts, each with different levels of complexity and interconnected hierarchically. The simplest model part (MP1) encapsulates only the dynamics related to LDH seen in
equation 5, without the presence of candidalysin. The next model part (MP2) integrates candidalysin as a damage-causing toxin as given in
equations 1 and
3-5, without the presence of the nanobody interaction. The more complex model part (MP3) incorporates blocking of candidalysin by the nanobody in
equation 2, and the final model part (MP4) estimates
C. albicans-related parameters by including
equations 6-8. A full overview of the parameter estimation procedure is depicted in
Fig. 6.
We fitted these models hierarchically using a bottom-up method, i.e., the estimates from the less complex model were used as fixed parameters for the more complex ones. Thus, we assumed that the nanobody and candidalysin interactions are comparable between synthetic and C. albicans-secreted candidalysin. Each model was fitted on a different data set vital for estimating the mechanisms of the associated data.
We employed the SciPy minimization package (
65) with 10 million different initial values for each fit to accurately pinpoint the true global minimum. The initial values were sampled using Latin Hypercube.
A profile likelihood method (
66) was used to assess the confidence interval for each parameter at a significance level of 0.05 surrounding the maximum-likelihood estimate. To evaluate the profile likelihood, we compared the chi-squared (χ
2) test statistics at the 95% percentile (approximately 3.84) with the negative logarithm of the likelihood ratio multiplied by 2. This approach enabled us to assess the practical identifiability of all parameters in our model. Figure S7 illustrates the profiles of the likelihood function for each parameter that was used to obtain the confidence intervals.
In silico model sensitivity analysis
A global Sobol sensitivity analysis was conducted using the open-source Python library SAlib (
67,
68) to assess the impact and the nature of the influence of relevant parameters. The parameters were tested for their first and total order Sobol sensitivity in relation to the amount LDH at various time points. The analysis involved sampling 4,096 points uniformly from the confidence intervals obtained through the profile likelihood method depicted in
Table 1 for each parameter.
In silico model numerical simulation
The numerical simulation of our model was executed using the LSODA solver incorporated in the SciPy library. LSODA is a versatile and robust solver adept at efficiently handling stiff systems of ordinary differential equations. The solver numerically integrated the system of equations over time, starting from the initial conditions of , and . The integration was carried out over a time range of interest, specifically [0, 72], with the two parameters atol for the absolute error and rtol for the relative error set to 1e-9 and 1e-10, respectively. The resulting time-dependent profiles of the key molecular entities in the system were then analyzed and compared to experimental measurements (refer to parameter estimation).