Fungal determinants contributing to translocation of Candida albicans yeast cells through the intestinal epithelial barrier

Abstract

Filamentous hyphae are the main invasive morphotype of the opportunistic fungal pathogen Candida albicans. However, yeast cells seem better suited for dissemination through the bloodstream during the progression of life-threatening systemic infections. While yeast cells are present together with hyphae in the intestine during commensal colonization, how yeast cells ultimately reach the blood following translocation of invasive hyphae is unknown. In this study we investigated potential mechanisms proposed for how yeast cells may enter the blood using an in vitro model of translocation based on intestinal epithelial cells (IECs). Our data show that yeast cells can passively translocate with invasive hyphae, though this requires host-cell damage facilitated by the peptide toxin candidalysin, encoded by ECE1. Independent of fungal-mediated damage, chemical disruption of the IEC layer by the mycotoxin patulin was sufficient to foster efficient translocation of C. albicans yeast cells alone. This was dependent on a significant loss of barrier integrity rather than host-cell damage itself. The same phenomenon was observed for oral clinical isolates, which more readily grow as yeast and pseudohyphal cells as compared to the standard SC5314 strain. The transition of hypha-to-yeast growth was also associated with translocation across IECs by increased expression of the yeast-essential gene PES1. This is the first study to directly investigate the mechanisms by which C. albicans yeast cells can translocate across IECs and to describe the fungal factors that contribute to this process.

Candida albicans hyphae following translocation across intestinal epithelial cells. The fungal microcolony shows GFP fluorescence under the control of the PES1 promoter, indicating lateral yeast cell formation.

Introduction

Invasive fungal infections are recognized as a serious threat to human health due to high mortality rates and few or ineffective treatment options (WHO 2022, Denning 2024). Among these, systemic candidiasis ranks as one of the most common life-threatening infections (Denning 2024, Lass-Flörl et al. 2024). Candida albicans is the species most often associated with such cases (Lass-Flörl et al. 2024). Normally existing as a commensal member of the microbiome in the gastrointestinal tract, this fungal population also serves as a reservoir for invasive infections (Kumamoto et al. 2020, Zhai et al. 2020, Rolling et al. 2021, Sprague et al. 2022).

Morphological plasticity has long been recognized as a key virulence factor and, more recently, an important commensal attribute for C. albicans (Jacobsen and Hube 2017, Schille et al. 2025). During commensal colonization of the murine gastrointestinal tract, the fungus grows as a mixed population of yeast and hyphae, though with varying proportions of each depending on the specific intestinal compartment (Witchley et al. 2019). After transitioning to a pathogenic stage in the intestinal niche, however, the hyphal cells are the main invasive morphotype of C. albicans (Goyer et al. 2016). In fact, non-filamentous strains of C. albicans are significantly reduced in their ability to translocate across intestinal epithelial cells (IECs) (Allert et al. 2018, Dunker et al. 2021, Sprague et al. 2024). Once in the bloodstream, the fungus is then able to disseminate and colonize distal organs leading to life-threatening systemic infections. The small size and shape of yeast cells originally led to the idea that they are better suited for dissemination compared to long, filamentous hyphae (Jacobsen et al. 2012). Indeed, yeast cells, especially those with short germ tubes, show an increased ability to adhere to endothelial cells as compared to hyphae or pseudohyphae in in vitro models with physiological flow (Grubb et al. 2009, Wilson and Hube 2010). Growth as yeast cells is also beneficial for dissemination through the bloodstream using in vivo mouse models of systemic candidiasis (Saville et al. 2003, Dunker et al. 2021).

If C. albicans invades and translocates across the intestinal epithelium as hyphae, but disseminates through the blood as yeast: how do yeast cells reach the bloodstream? For patients with underlying medical conditions like intestinal surgery or chemotherapy, damage to the intestinal barrier provides a clear route for yeast cells to directly reach the blood from the intestine (Eggimann and Pittet 2014, Eggimann et al. 2015, Böhm et al. 2017, McDonough et al. 2021). Additionally, intravenous medical devices contaminated with C. albicans biofilms can serve as a source for systemic infections. However, many patients at risk for developing systemic candidiasis are critically ill without obvious signs of mucosal damage to the intestine or contact with contaminated devices (Eggimann and Pittet 2014, Eggimann et al. 2015, Davidson et al. 2018). In these susceptible patients, it remains unclear how C. albicans yeast cells reach the bloodstream. A number of hypotheses to answer this question have been proposed, namely mechanical lesions, a “hitchhiking” mechanism, and a hypha-to-yeast transition (Sprague et al. 2022). As mentioned above, many patients at a high risk for developing systemic candidiasis present with a disturbed intestinal epithelium, either directly from surgery or due to underlying diseases and treatments like chemotherapy (Eggimann and Pittet 2014, Eggimann et al. 2015, Davidson et al. 2018). In these cases, mechanical lesions present a probable route for yeast cells to enter the bloodstream. However, damage to the intestinal epithelium can also come directly from the fungus, especially given that there is an expansion of Candida spp. in the intestine prior to systemic infections (Zaborin et al. 2014, Zhai et al. 2020). In this case, C. albicans could cause enough epithelial damage itself to allow yeast cells to translocate along with invading hyphae via a so-called hitchhiking mechanism, either by transiently attaching to the hyphae or passively moving through the resulting lesions (Sprague et al. 2022). Finally, there is the possibility that after translocation to the bloodstream, C. albicans hyphae undergo another morphological transition to yeast cells, the so-called hypha-to-yeast transition.

In this study, we sought to discern whether each of these proposed mechanisms contribute to translocation of C. albicans across the intestinal epithelial barrier in vitro. Host-cell damage mediated by candidalysin, encoded by the gene ECE1 and produced by invading hyphae, allows passive translocation of yeast cells across IECs, possibly via a hitchhiking mechanism. However, fungal damage potential and filamentation are not required for efficient translocation when the intestinal barrier integrity is disrupted. We show that expression of the fungal PES1 gene is triggered in hyphae following translocation across, but not during invasion of IECs. Altogether, we found that all three mechanisms likely contribute to fungal translocation depending on the specific predisposing conditions and the resulting effect on the intestinal epithelium.

Materials and methods

Candida albicans strains and growth conditions

Candida albicans strains used in this study are shown in Table 1. Strains were routinely cultivated on/in YPD agar/broth (1% yeast extract, 2% peptone, 2% d-glucose with or without 1.5% agar) at 30°C. Overnight (O/N) cultures were cultured for 16 h in YPD broth at 30°C with shaking at 180 rpm unless otherwise specified. Cells were then washed twice with phosphate-buffered saline (PBS) and the cell number was adjusted.

Table 1.

Candida albicans strains used in this study.

Strain	Description	Source
SC5314	Wild type, bloodstream isolate	Gillum et al. (1984)
ece1ΔΔ	SC5314 background, ece1::∆/ece1::∆ NEUT5L/neut5l::FRT	Weerasinghe et al. (2025)
efg1ΔΔcph1ΔΔ	SC5314 background, efg1::FRT/efg1::FRT cph1::FRT/cph1::FRT	Wartenberg et al. (2014)
SC5314 RFP	SC5314 background, ADH1/adh1::(tetO-POP4-dTomato, SAT1)	This study
ece1ΔΔ RFP	SC5314 background, ece1::∆/ece1::∆ NEUT5L/neut5l::FRT ADH1/adh1::(tetO-POP4-dTomato, SAT1)	This study
C73	OPC isolate	Received from Julian Naglik
529L	OPC isolate	Rahman et al. (2007)
PEV-GFP	SC5314 background, RPS1/rps1::(yeGFP, SAT1)	This study
PPES1-GFP	SC5314 background, RPS1/rps1::(PPES1 -yeGFP, SAT1)	This study
SC5314 mScarlet	SC5314 background, ADH1/adh1::(tetO-POP4-mScarlet, SAT1)	Valentine et al. (2025)

Culture of epithelial cells

The intestinal epithelial Caco-2 brush border expressing 1 cell line (C2BBe1; ATCC, CRL2102) (Peterson and Mooseker 1992) and the human intestinal goblet cell line (HT29-MTX; ATCC, HTB-38; CLS, Lot No. 13B021) were routinely cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) (Bio&Sell), 10 µg/ml Holotransferrin (Calbiochem, Merck), and 1% non-essential amino acids (Gibco, Thermo Fisher Scientific) at 37°C with 5% CO₂ for no longer than 15 passages. C2BBe1 and HT29-MTX cells were seeded at a 70:30 ratio (C2BBe1:HT29-MTX) and differentiated for 12 d at 37°C with 5% CO₂ and regular medium exchange after reaching 100% confluency. The human oral epithelial cells (TR146, European Collection of Authenticated Cell Cultures ECACC #10 032 305) were cultivated at 37°C and 5% CO₂ in Dulbecco’s modified Eagle medium/Nutrient Mixture F-12 (DMEM/F12, Life Technologies) supplemented with 10% fetal bovine serum (FBS) (Bio&Sell) for no longer than 15 passages. TR146 cells were seeded and incubated for 2 d at 37°C with 5% CO₂ until the cells reached confluency. For damage (C2BBe1: HT29-MTX or TR146) and translocation (C2BBe1: HT29-MTX) assays, a total of 2 × 10⁴ cells per well of a 96-well plate or per transwell insert (polycarbonate membrane with 5 μm pores; Corning) were seeded, respectively. For filamentation assays, C2BBe1: HT29-MTX cells or TR146 cells were seeded in 24-well plates with glass coverslips at a total concentration of 1 × 10⁵ cells per well. All well plates, transwell inserts, and coverslips for experiments with IECs were coated with collagen I (10 μg/ml for 2 h at room temperature; Thermo Fisher Scientific) prior to seeding. Just prior to infection with C. albicans, the medium was removed, and fresh DMEM or DMEM/F12 without any supplementation was added to the cells.

In vitro translocation assay

Differentiated C2BBe1: HT29-MTX cells in transwell inserts were infected with 1 × 10⁵  C. albicans cells and incubated at 37°C with 5% CO₂. Epithelial barrier integrity was determined by measuring the trans-epithelial electrical resistance (TEER) using a volt-ohm meter (EVOM2, World Precision Instruments). Supernatant from the upper compartment was removed after 24 h and used for LDH measurement as described below. The lower compartment was treated with 20 U/ml zymolyase (Amsbio) for 2 h at 37°C and 5% CO₂ to detach translocated hyphae. The zymolyase-treated hyphae were then plated on YPD agar, incubated at 30°C for 2 d, and the colony forming units (CFUs) were counted. For experiments where IECs were infected with mixed C. albicans strains, IECs were infected with 1:1 mixtures of two C. albicans strains at a final concentration of 1 × 10⁵ fungal cells per transwell. The combinations were SC5314: ece1ΔΔ-RFP, SC5314-RFP: efg1ΔΔcph1ΔΔ, and ece1ΔΔ-RFP: efg1ΔΔcph1ΔΔ. Colonies from the different strains were differentiated by the red color of RFP-expressing strains. For experiments with patulin or dextran sodium sulfate (DSS) treatment, IECs were pre-treated for 1 h at 37°C with 5% CO₂ with either patulin (10 or 25 μM) or DSS (3%) in DMEM in the lower compartment. The IECs were then infected with C. albicans and incubated as described above. Patulin was originally dissolved in dimethyl sulfoxide (DMSO); therefore the mock-treated samples were treated with diluted DMSO.

The sensitivity of the two oral C. albicans isolates (C73 and 529L) to treatment with zymolyase was tested as described previously (Allert et al. 2018). Briefly, 1 × 10⁵  C. albicans cells were seeded into a 96-well plate in DMEM and incubated for 3 h at 37°C with 5% CO₂ to induce hyphae. Zymolyase was added to a final concentration of 20 U/ml and incubated for a further 2 h at 37°C with 5% CO₂. The fungal cells were then diluted, plated on YPD agar, and fungal colonies were counted after incubation at 30°C for 2 d. Both C73 and 529L were not significantly more sensitive or resistant to the zymolyase treatment used during translocation assays compared to the SC5314 wild type strain (SI Fig. 2).

Quantification of host cytotoxicity (LDH assay)

Differentiated C2BBe1: HT29-MTX cells or TR146 in 96-well plates were infected with C. albicans cells and incubated at 37°C with 5% CO₂ for 24 and 48 h. IECs were infected with a multiplicity of infection (MOI) of 4 and OECs were infected with an MOI of 1. Epithelial damage was quantified by release of lactate dehydrogenase (LDH) in supernatants from IECs and OECs in 96-well plates as well as IECs from transwell inserts. LDH concentrations were measured using a cytotoxicity detection kit (Roche) according to the manufacturer’s instructions and LDH from rabbit muscle (Roche) was used to generate a standard curve. The data are shown as % LDH release from uninfected/mock-treated epithelial cells.

Quantification of filament length and morphology

Intestinal or oral cells in a 24-well plate were infected with C. albicans at a MOI of 1 and incubated at 37°C with 5% CO₂ for 3 h. Non-adherent C. albicans cells were removed by rinsing with PBS and samples were fixed with Roti®-Histofix 4% (Roth). The fungal cells were then stained with Calcofluor white (CFW; Sigma-Aldrich) for 30 min at room temperature and visualized in the inverted Axio Observer.Z1 microscope (Zeiss) using a 40x/0.95-numerical-aperture air objective (Zeiss) in bright field and excitation at 385 ± 15 nm and emission at 425 ± 13 nm. The percentage of filaments, pseudohyphae, and yeast was counted, and the total filament length was measured from at least 100 hyphae per strain for each biological replicate in ZEN 3.10 (Zeiss).

Candida albicans strain construction

The pNIM1R-dTOM2 plasmid (Prieto et al. 2014) was used to generate RFP-labeled C. albicans strains. The transformation cassette was excised with SacII and KpnI and transformed into SC5314 or ece1ΔΔ and selected for on YPD medium containing 200 µg/ml nourseothricin. Integration into the ADH1 locus was confirmed by colony PCR.

Transcriptional fluorescent reporters were generated with the pGFP vector (Barelle et al. 2004). The URA3 selection marker of pGFP was replaced with the SAT1 marker. This plasmid contains a promoterless, yeast-optimized GFP gene and served as the empty vector (EV) control. The promoter of PES1 was amplified from SC5314 genomic DNA and cloned into pGFP at XhoI/HindIII. The transformation cassettes were linearized by digestion with StuI, transformed into SC5314, and selected for on YPD medium containing 200 µg/ml nourseothricin. Integration into the RPS1 locus was confirmed by colony PCR. The primers and their corresponding sequences are given in Table 2.

Table 2.

Primers used in this study.

Name	Sequence 5’ to 3’
SAT1-F	ACCGCGGTGGCGGCCGCAGCGTCAAAACTAGAGAATAATAAA
SAT1-R	ATCCACTAGTTCTAGAAGGACCACCTTTGATTGTAAATAGT
PES1p-XhoI-F	CGGGCCCCCCCTCGAGGCAGCAAGAACCCAAGTGAC
PES1p-HindIII-R	CATTTTAATAAAGCTTTTTATGTTTGTTTGGTTTGATGGAG

Microscopy and image analysis of translocated hyphae

Differentiated C2BBe1: HT29-MTX cells in transwell inserts were infected with 5 × 10⁴ cells of an mScarlet-expressing C. albicans strain. After incubation at 37°C for 24 h with 5% CO₂, the transwell membranes were removed and immediately mounted on glass microscopy slides. Translocating fungi were visualized in the inverted Axio Observer.Z1 microscope (Zeiss) using a 63x/1.4-numerical-aperture oil immersion objective (Zeiss) in bright field and excitation at 574 ± 12 nm and emission at 605 ± 35 nm. Orientation of the membrane was determined based on the presence of microcolonies on the translocated side, visible membrane pores, and host cells as well as a fungal material on the epithelial side. To determine PES1 expression in translocating fungal material, differentiated C2BBe1: HT29-MTX cells in transwell inserts were infected with a C. albicans strain expressing green fluorescent protein (GFP) under the control of the PES1 promoter. After 24 h, the transwell membranes were removed and immediately mounted on glass microscopy slides. Afterwards, the samples were imaged in the same microscope with a 20x/0.8-numerical-aperture air objective (Zeiss) in bright field and excitation at 470 ± 20 nm and emission at 525 ± 25 nm. At least three microcolonies of at least two technical replicates per biological replicate were imaged with the same settings. For the z-stacks, at least two microcolonies per biological replicate were imaged with the same settings. Z-positions were adjusted to include the translocated microcolony and the fungal material on the epithelial side. Afterwards, raw images were exported as .tif using ZEN 3.10 (Zeiss) and imported into ImageJ. Using the freehand selection tool, outlines of the microcolony or fungal material were drawn in the brightfield image and stored as a ROI in the ROI Manager. The ROI was then added to the corresponding raw GFP-channel image and mean fluorescence intensity was measured. The measured intensities were saved as a .csv and imported into Excel (Microsoft). Finally, the mean fluorescence intensity and the ratio of intensity between the translocated microcolony and the fungal material on the epithelial side were calculated for each biological replicate.

For the visualization, representative images were exported as .tif with the same display settings using ZEN 3.10 (Zeiss) for all strains.

Statistical analyses

All experiments were performed with at least three biological replicates. Data were analyzed using Prism 10.4.0 (GraphPad Software).

Results

Candida albicans yeast cells can “hitchhike” with invading hyphae to translocate across IECs

To decipher the mechanisms that contribute to yeast entering the bloodstream, we first investigated whether invading C. albicans hyphae can serve to deliver colonizing yeast cells across the epithelial barrier. Intestinal epithelial cells (IECs) were infected with different strains of C. albicans and incubated for 24 h, when the barrier integrity was determined by measuring the trans-epithelial electrical resistance (TEER), host-cell damage was measured as a release of the cytoplasmic enzyme lactate dehydrogenase (LDH), and fungal translocation was measured as the number of translocated colony-forming units (CFUs). When infected individually, a strain lacking the candidalysin-encoding gene ECE1 (ece1ΔΔ) (Lortal et al. 2025) translocated on average less than half the rate of the parental wild type (WT) (SC5314) strain, while a non-filamentous strain (efg1ΔΔcph1ΔΔ) had an even further reduced ability to translocate (Fig. 1A). However, when the strains were infected together at a 1:1 ratio, only the combination of the hypha- and damage-competent WT strain with the non-filamentous efg1ΔΔcph1ΔΔ strain increased the translocation of the yeast-locked strain (Fig. 1B). This was accompanied with a significant decrease in the barrier integrity of the IECs infected with WT: efg1ΔΔcph1ΔΔ compared to the efg1ΔΔcph1ΔΔ alone (Fig. 1C). While the barrier integrity of IECs infected with ece1ΔΔ: efg1ΔΔcph1ΔΔ was similar to those infected with WT: efg1ΔΔcph1ΔΔ, this was not a significant decrease compared to IECs infected with efg1ΔΔcph1ΔΔ alone. This phenotype was not observed when the efg1ΔΔcph1ΔΔ strain was combined with the ece1ΔΔ strain. There was no significant effect when the WT and ece1ΔΔ strains were infected together, compared to infection with either WT or the ece1ΔΔ strains alone (Fig. 1C). This suggests that host-cell damage caused by invading C. albicans hyphae to IECs is sufficient to allow passive translocation of yeast cells across the intestinal barrier, possibly via a hitchhiking mechanism. However, whether this also involves direct binding of the yeast cells to the invading hypha is not clear.

Figure 1.

Co-infection with WT C. albicans, but not a non-damaging strain, fosters translocation of C. albicans yeast cells. (A) Translocation of individual C. albicans strains in a mixed C2BBe1: HT29-MTX. The non-filamentous strain lacking EFG1 and CPH1 is unable to translocate across IECs, while the non-damaging strain lacking ECE1 is severely impaired. (B) Translocation of C. albicans strains infected in combination at a 1:1 ratio where one of the strains constitutively expressed RFP. Co-infection of the efg1ΔΔcph1ΔΔ strain with WT cells increased the translocation rate, while co-infection with the non-damaging ece1ΔΔ strain had no effect. (C) Epithelial barrier integrity as measured by TEER during single and co-infections with C. albicans strains. Co-infection of the efg1ΔΔcph1ΔΔ strain with both WT or the ece1Δ/Δ strain decreased the barrier integrity compared to infection with the efg1ΔΔcph1ΔΔ strain alone. Data were compared with a 1-way ANOVA. Translocation CFU data (A) were tested for significance with a post-hoc Dunnett’s test for multiple comparisons and barrier integrity TEER data (C) were tested for significance with a post-hoc Šidák’s test for multiple comparisons. *, P ≤ 0.05; **, P ≤ 0.01; ^****, P ≤ 0.0001.

Loss of epithelial barrier integrity is required for passive translocation of C. albicans yeast cells

While hypha formation and candidalysin-mediated damage are required for efficient intestinal translocation in vitro, many patients susceptible to systemic candidiasis already have predisposing factors that result in damage to the intestinal epithelium (Eggimann and Pittet 2014, Eggimann et al. 2015, Allert et al. 2018, Davidson et al. 2018, Sprague et al. 2024). In these cases, is the existing damage to the epithelial barrier sufficient to allow passive translocation by colonizing yeast cells or is additional host-cell damage mediated by the fungus also required? To investigate this, IEC layers were chemically disturbed with the mycotoxin patulin (Kawauchiya et al. 2011, Goyer et al. 2016, Akbari et al. 2017). Patulin induces increased barrier permeability and a loss of TEER in a concentration- and time-dependent manner (Akbari et al. 2017). At a concentration of 10 μM, there was no significant effect on LDH release of IECs after 24 h, even during C. albicans infection (Fig. 2A). Despite this, the barrier integrity was significantly lower upon treatment with 10 μM patulin under all conditions; however, this did not translate to a significant increase in fungal translocation for any strain (Fig. 2B and C). Though there was a more than 70-fold increase in translocated CFUs for the efg1ΔΔcph1ΔΔ strain, it was not statistically significant (Fig. 2C). Treatment with patulin at a higher concentration (25 μM) was able to further decrease the barrier integrity of the IEC layers while also significantly increasing LDH release even for uninfected cells (Fig. 2A and B). Under these conditions where the epithelial layer was significantly damaged and barrier integrity was severely reduced, translocation of the efg1ΔΔcph1ΔΔ strain increased drastically and even surpassed that of the SC5314 and ece1ΔΔ strains (Fig. 2C). However, the fact that the translocated CFUs for the efg1ΔΔcph1ΔΔ strain are nearly 10-fold higher than the SC5314 and ece1ΔΔ strains may also be due in part to the differences in morphology of the tested strains.

Figure 2.

Passive translocation of C. albicans yeast cells requires severe disruption to the intestinal epithelial barrier integrity. (A) Treatment of IECs with 10 μM patulin does not significantly affect C. albicans-mediated damage; however, treatment with 25 μM significantly increases host-cell damage even in the absence of any fungus. Data are presented as the percentage of LDH released by uninfected, mock-treated IECs. (B) Patulin treatment of IECs significantly decreases the barrier integrity of the epithelial layer at 10 and 25 μM, regardless of whether the host cells were infected with C. albicans. (C) Fungal translocation increases for all fungal strains upon treatment of IECs with patulin; however, this is only to a statistically significant degree for the efg1ΔΔcph1ΔΔ strain during treatment with 25 μM patulin. Six biological replicates were done for mock-treated samples for each strain. Values of zero within the dataset were included for performing statistical analyses, but cannot be plotted on a logarithmic scale. Data were compared with a 2-way ANOVA and tested for significance with a post-hoc Tukey’s test for multiple comparisons. (D) Treatment of IECs with 3% DSS significantly increases LDH release regardless of C. albicans infection. (E) DSS treatment of IECs significantly decreased barrier integrity for all samples, though not to the same degree. (F) Fungal translocation increased for all strains upon treatment of IECs with DSS, though only to a statistically significant degree for the ece1ΔΔ strain, which had nearly double the CFUs compared to the mock-treated samples. For the patulin-treated, uninfected (A) and DSS-treated, uninfected (D) LDH samples, significance was determined using one-sample t-tests with comparison to a theoretical value of 100. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ^****, P ≤ 0.0001.

To confirm the correlation with passive translocation of C. albicans yeast cells and disrupted intestinal barrier integrity, the same experiments were performed using dextran sodium sulfate (DSS). DSS is commonly used in vivo in murine models to induce colitis and has been used in systemic models of hematogenously disseminated candidiasis to induce fungal translocation to the bloodstream from the intestine (Koh et al. 2008, Yang and Merlin 2024). Treatment with 3% of DSS in cell culture medium was sufficient to significantly increase the LDH release of infected and uninfected IECs (Fig. 2D). This also significantly decreased the barrier integrity for all conditions tested; however, not to the same degree as treatment with 25 μM patulin (Fig. 2B and E). Despite the increased LDH release and decreased barrier integrity, treatment with DSS was not able to significantly increase the fungal translocation for the WT or non-filamentous efg1ΔΔcph1ΔΔ strain (Fig. 2F). The translocation of the ece1ΔΔ strain was significantly increased upon DSS treatment and reached levels of SC5314. These data suggest that increased cellular damage of IECs per se does not increase the ability of C. albicans yeast to passively translocate across the epithelial barrier, but rather only when it results in significant disruption of the overall barrier integrity independent of the fungus’ ability to induce epithelial damage. Additionally, increased host-cell damage and decreased barrier integrity are sufficient to rescue the translocation defect of the hypha-competent but non-damaging ece1ΔΔ strain.

Oral clinical isolates of C. albicans only translocate efficiently with a disrupted intestinal epithelial barrier

In recent years, it has become apparent that the standard WT (SC5314) does not accurately represent the majority of C. albicans isolates (MacCallum et al. 2009, Iracane et al. 2024, Zhang et al. 2025). In fact, many clinical isolates of C. albicans are reduced in their capacity to form hyphae and damage host cells in vitro compared to SC5314 while outcompeting the standard WT strain in colonization of the gastrointestinal tract of antibiotic-treated mice (MacCallum et al. 2009, Liu et al. 2021, McDonough et al. 2021, Zhang et al. 2025). As our data show that non-filamentous C. albicans strains can translocate across a damaged intestinal epithelial barrier (Fig. 2C and F), we further investigated whether the same effect was true for such clinical isolates. Two clinical strains isolated from the oral cavity (C73 and 529L) were selected, which showed a diminished ability to damage oral epithelial cells (OECs) even after 48 h of infection (SI Fig. 1A). This could be attributed to decreased hypha formation and an increased propensity for yeast or pseudohyphal growth during infection for each strain (SI Fig. 1B and C). Both strains were similarly unable to cause a significant release of LDH from IECs during infection after either 24 or 48 h compared to the standard lab WT strain, SC5314 (Fig. 3A). Similar to infection of OECs, both oral isolates also showed decreased filament length on IECs (Fig. 3B) as well as an increased tendency to grow as either pseudohyphae or yeast cells (Fig. 3C and D). Due to the impaired ability of C73 and 529L to damage epithelial cells and their decreased hypha formation, we hypothesized that substantial damage to the intestinal epithelium would be required to allow efficient translocation for such C. albicans strains.

Figure 3.

Clinical isolates from the oral niche do not readily form hyphae and have a drastically reduced damage potential during in vitro infection. (A) Both oral isolates, C73 and 529L, elicit significantly less LDH release from IECs after 24 or 48 h of infection as compared to the SC5314 WT strain. (B) While SC5314 shows robust filamentous growth during infection of IECs, both C73 and 529L have significantly decreased filament length after 3 h. (C) During infection of IECs, both C73 and 529L more readily grow as pseudohyphae and yeast as opposed to hyphae compared to the SC5314 strain. Asterisks depict statistically significant differences in the percentage of hyphae and pseudohyphae for both C73 and 529L compared to the SC5314 strain, but not for yeast cells (P=0.4448 for C73; P=0.1594 for 529L). (D) Representative images of each C. albicans strain after 3 h of infection of IECs. Samples were fixed and C. albicans was stained with calcofluor white. Scale bars are set to 20 μm. LDH release (A) and morphology data (C) were compared with a 2-way ANOVA and tested for significance with a post-hoc Tukey’s test for multiple comparisons. Filament length data (B) were compared with a 1-way ANOVA and tested for significance with a post-hoc Dunnett’s test for multiple comparisons. **, P ≤ 0.01; ^****, P ≤ 0.0001.

Unsurprisingly, due to their inability to induce host-cell damage, both oral isolates were unable to significantly decrease the barrier integrity or efficiently translocate during infection of IECs (Fig. 4A–C). Treatment of IECs with 10 μM patulin did not have an effect on the LDH release during infection with the oral isolates; however, treatment with 25 μM significantly increased LDH release to a similar degree in the uninfected samples (Fig. 4A). Despite no change in LDH release, treatment with 10 μM patulin was sufficient to significantly decrease the barrier integrity during infection of IECs with both oral isolates; however, the TEER values again did not change more than in the uninfected samples (Fig. 4B). Similar to the data above (Fig. 2B), treatment with 25 μM patulin led to a further decrease of the barrier integrity under all infection conditions (Fig. 4B). Fungal translocation of the oral isolates also only significantly increased upon treatment with 25 μM patulin, similar to the efg1ΔΔcph1ΔΔ strain (Fig. 4C). While translocation of the efg1ΔΔcph1ΔΔ strain increased nearly 20 000-fold (Fig. 2C), the increases were around 300-fold and 280-fold for C73 and 529L upon patulin treatment, respectively (Fig. 4C). While these are significant changes in the translocation rate for both oral isolates, they still do not reach the same degree as the translocation rate of SC5314 without any patulin treatment.

Figure 4.

Efficient translocation of hypovirulent oral C. albicans isolates requires severe disruption to the intestinal epithelial barrier integrity. (A) Treatment of IECs with 10 μM patulin has no significant impact on host-cell damage for the additional clinical isolates C73 and 529L; however, treatment with 25 μM patulin increases LDH release, though only to a statistically significant degree for the C73 strain. (B) Patulin treatment of IECs significantly decreases the barrier integrity of the epithelial layer at 10 and 25 μM, regardless of whether the host cells were infected with C. albicans. (C) Fungal translocation increases for both oral C. albicans isolates upon treatment with both 10 and 25 μM patulin, nearly reaching the rate of the SC5314 strain upon treatment with 25 μM. Six biological replicates were done for mock-treated samples for each strain. Values of zero within the dataset were included for performing statistical analyses, but cannot be plotted on a logarithmic scale. (D) Treatment of IECs with 3% DSS significantly increases LDH release during infection with both oral isolates. (E) DSS treatment of IECs significantly decreased barrier integrity for all samples, though only to just below that of the mock-treated control for samples infected with the oral isolates. (F) Fungal translocation for the C. albicans oral isolates was not impacted by treatment of the IECs with DSS. Three biological replicates were done for mock-treated samples for each strain. Values of zero within the dataset were included for performing statistical analyses, but cannot be plotted on a logarithmic scale. Data were compared with a 2-way ANOVA and tested for significance with a post-hoc Tukey’s test for multiple comparisons. For the patulin-treated, uninfected (A) and DSS-treated, uninfected (D) LDH samples, significance was determined using one-sample t-tests with comparison to a theoretical value of 100. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ^****, P ≤ 0.0001.

Disruption of the intestinal barrier with DSS treatment was similarly able to significantly increase the LDH release of IECs during infection with C73 and 529L (Fig. 4D). However, similar to our data with the efg1ΔΔcph1ΔΔ strain (Fig. 2E), treatment with DSS significantly reduced the barrier integrity during infection with both oral isolates, but not to the same degree as treatment with 25 μM patulin (Fig. 4B and E). The increased cellular damage of IECs without a severe decrease in barrier integrity was not itself sufficient to significantly increase fungal translocation for the oral isolates. Both C73 and 529L showed no significant change to translocation rates upon treatment of IECs with DSS (Fig. 4F). These data further suggest that increased cellular damage to IECs alone is not sufficient to foster efficient translocation of non-damaging and hypofilamentous C. albicans strains.

Translocated hyphae undergo a hypha-to-yeast transition

While the non-filamentous efg1ΔΔcph1ΔΔ strain was able to hitchhike with invading hyphae of the SC5314 WT strain, there were still nearly 6-fold more translocated CFUs from the SC5314 strain (Fig. 1B). While significant damage to the epithelial barrier was sufficient to allow efficient passive translocation of yeast (Fig. 2C), less severe disruption of the barrier was not itself sufficient to have the same effect (Fig. 2C and F). Additionally, not all critically ill patients at risk to develop systemic candidiasis present with obvious signs of damage to the intestinal epithelium (Eggimann and Pittet 2014, Eggimann et al. 2015, Davidson et al. 2018). Another hypothesis for how C. albicans yeast cells reach the bloodstream is a transition from filamentous, hyphal growth back to budding yeast (Sprague et al. 2022). Using an mScarlet-expressing C. albicans SC5314 strain, translocating hyphae were imaged after 24 h of infection of IECs. Most fungal cells on the underside of the porous membrane were hyphae present in microcolonies (Fig. 5A). Despite this, there were visible yeast cells attached to some of the translocated hyphae. While the membranes were mounted on slides directly after being removed and imaged immediately, some attached yeast cells and those that had already detached from the hyphae were likely lost during preparation and removal of the medium from the samples.

Figure 5.

Candida albicans hyphae express PES1 and form lateral yeast cells following translocation across IECs. (A) Microscopy image of an mScarlet-expressing C. albicans microcolony on the underside of a transwell membrane following translocation across an IEC layer. The white arrow indicates lateral yeast formed from hyphae. (B) Quantification of mean fluorescence intensity in translocated microcolonies. The P_PES1-GFP transcriptional reporter shows a significantly higher fluorescence signal in translocated hyphae compared to the P_EV-GFP control strain. (C) Fold change of mean fluorescent signal of translocated microcolonies to those associated with IECs for P_EV-GFP and P_PES1-GFP. The P_PES1-GFP reporter strain shows a significantly higher fold change in the mean fluorescence intensity of translocated microcolonies relative to hyphae associated with the IEC layer compared to the control P_EV-GFP strain. (D) Representative images of translocated microcolonies of the P_PES1-GFP transcriptional reporter and the corresponding P_EV-GFP control strain. White outlines indicate the regions of interest (ROI) used for quantification of the fluorescent signal. (E) Representative images of P_EV-GFP and P_PES1-GFP microcolonies associated with IECs (left column) and after translocation (right column). White outlines indicate the ROIs used for quantification of the fluorescent signal. Membranes were imaged along the Z axis to visualize translocated microcolonies and epithelial fungal material within the same sample. The entire Z-stack of the representative P_EV-GFP and P_PES1-GFP samples are shown in SI Videos 1 and 2, respectively. Scale bars are set to 20 μm. Mean fluorescence intensity (B) and fold change (C) data were compared using an unpaired t-test. *, P ≤ 0.05.

To overcome this, expression of the fungal PES1 gene was visualized in translocating hyphae. The PES1 gene in C. albicans is essential for yeast growth and required for the formation of lateral yeast cells from hyphae as well as disperser cell formation from biofilms (Uppuluri et al. 2012, Uppuluri et al. 2018). Therefore, a PES1 transcriptional reporter strain was constructed with a yeast-optimized GFP gene under the control of the PES1 promoter and a corresponding control plasmid with a promoterless GFP (P_EV-GFP) (Table 1). During infection of IECs, hyphae that had translocated across the epithelial barrier showed a stronger fluorescent signal in the P_PES1-GFP reporter strain as compared to the control P_EV-GFP strain (Fig. 5D). The mean fluorescent signal of regions surrounding translocated fungal microcolonies was quantified for P_PES1-GFP and P_EV-GFP. There was a significantly higher fluorescence intensity in the PES1 transcriptional reporter strain compared to the control strain (Fig. 5B). Additionally, the fluorescent signal was significantly stronger in translocated hyphae on the underside of the transwell membrane compared to hyphae still associated with the IEC layer on the upper side of the membrane (Fig. 5E; SI Videos 1 and 2). The change in mean fluorescence intensity in translocated microcolonies relative to those associated with the epithelial layer was compared between the two strains. The difference in fluorescence intensity between translocated hyphae and IEC-associated hyphae was significantly higher for the P_PES1-GFP strain compared to the P_EV-GFP strain (Fig. 5C). These data suggest that following translocation across the intestinal epithelium, C. albicans hyphae show increased expression of PES1, which may indicate or even trigger a hypha-to-yeast transition prior to dissemination through the bloodstream.

Discussion

In this study, we explored different routes for C. albicans yeast cells to reach the bloodstream from the intestinal tract at the onset of systemic infection. Using an in vitro model of intestinal translocation, we show that yeast cells are able to hitchhike with invading hyphae, that chemical disruption of the epithelial barrier allows for passive translocation of yeast cells, and that hyphae form lateral yeast following translocation. We also dissected the contributions of hypha formation, host-cell damage, and integrity of the intestinal barrier to fungal translocation for less virulent clinical isolates of C. albicans.

While C. albicans hyphae are the main invasive morphotype, yeast cells with short germ tubes are optimal for adhesion to endothelial cells under flow and thus better suited for dissemination through the bloodstream (Grubb et al. 2009, Wilson and Hube 2010, Jacobsen et al. 2012). It has been hypothesized that yeast cells present in the intestine can hitchhike together with invading hyphae in order to reach the bloodstream and disseminate to distal organs at the onset of systemic Candida infections (Sprague et al. 2022). This could occur by yeast cells passively translocating following significant epithelial damage mediated by the invading hyphae. Our study provides evidence to support this, as a non-filamentous strain of C. albicans was unable to cross an intact layer of IEC alone, but the translocation efficiency was significantly increased upon co-infection with a hypha-competent, damaging WT strain. This suggests that in a mixed population of yeast and hyphal cells on the intestinal barrier, yeast cells can translocate along with invading hyphae. However, co-infection with a C. albicans strain lacking the ECE1 gene was not able to increase the translocation of a non-filamentous strain. This was expected, as host-cell damage mediated by ECE1/candidalysin is required for efficient translocation of C. albicans across IECs in vitro (Allert et al. 2018). Though host-cell damage mediated by the fungus is not a consistent predictor for fungal translocation. During infection with C. albicans in vitro, NFκB signaling in IECs maintains barrier integrity and prevents fungal-mediated damage and translocation (Sprague et al. 2024). However, upon inhibition of NFκB activation, even a non-filamentous C. albicans strain translocates efficiently in the absence of a significant increase in host-cell damage. Therefore, when the host epithelial response is impaired, this can lead to loss of barrier integrity and ultimately fungal translocation. While co-infection of the non-filamentous strain with the WT significantly decreased the barrier integrity of IECs, co-infection with the strain lacking ECE1 did not. Thus, our data show that invading hyphae that induce enough host-cell damage to decrease the epithelial barrier integrity can foster translocation of yeast cells present in the intestine. However, it is still unclear whether this process involved transient attachment of the yeast cells to the surface of the invading hyphae.

Many patients at risk for and who ultimately develop systemic candidiasis suffer from predisposing conditions that compromise their intestinal epithelial barrier (Eggimann and Pittet 2014, Eggimann et al. 2015, Böhm et al. 2017, McDonough et al. 2021). In this context, it was unclear whether the pre-existing damage to the intestinal barrier is sufficient to allow fungal translocation or whether fungal-mediated damage would still be required. Mouse models of systemic C. albicans infection arising from the intestine require antibiotic treatment for fungal colonization and overgrowth, neutropenia to weaken the host immune system, and mucosal damage to the intestine to induce fungal translocation (Koh et al. 2008). Our data suggest that only severe disruption of the intestinal epithelial barrier can induce efficient translocation of non-damaging or non-filamentous C. albicans strains. We show that disruption of IEC layers with the mycotoxin patulin could only induce translocation of a non-filamentous C. albicans strain similar to that of the WT at concentrations where host-cell damage was significantly increased and epithelial barrier integrity was drastically decreased. However, host-cell damage and loss of epithelial barrier integrity, though often associated with one another, can be independent processes (Böhringer et al. 2016, Allert et al. 2018, Sprague et al. 2024). In line with this, treatment of IECs with DSS significantly increased host-cell damage similar to the highest concentration of patulin tested. However, while treatment with this concentration of patulin drastically reduced the TEER to similar levels for all conditions tested, this was not the case for IECs treated with DSS. IECs treated with DSS rather showed a moderate, though still significant loss of barrier integrity despite the increased host-cell damage. However, this was still unable to allow for efficient translocation of a non-filamentous C. albicans strain. Taken together with previous studies, our data further indicate that disruption of the epithelial barrier is necessary for fungal translocation from the intestine and that host-cell damage fosters translocation indirectly via the resulting decreased barrier integrity. In line with this, severe loss of barrier integrity, whether fungal-induced or due to underlying host conditions, is necessary to allow for passive translocation of yeast cells from the intestine.

Our observations indicate that sufficient loss of intestinal barrier integrity is a pre-requisite for passive translocation of yeast cells and that invading hyphae can further increase this efficiency. However, new studies continue to highlight the genetic and phenotypic variability between different isolates of C. albicans (MacCallum et al. 2009, Anderson et al. 2023, Iracane et al. 2024, Zhang et al. 2025). Indeed, many clinical isolates show varying degrees of filamentation and virulence potential in vitro. Our data fit well with these observations, as both oral C. albicans isolates tested showed decreased hypha formation and damage potential during in vitro epithelial infection. We found that both oral isolates behaved similarly to the non-filamentous strain during in vitro IEC infection. Disturbance of the epithelial barrier with patulin could only significantly increase the translocation efficiency of the oral isolates at concentrations high enough to increase host-cell damage and severely reduce barrier integrity. Additionally, both oral isolates showed no change in translocation upon treatment with DSS where there was a significant increase in host-cell damage with only a modest decrease in barrier integrity. While a similar trend as with the non-filamentous strain was seen for both oral isolates, the change in translocation efficiency was not as drastic. This could be due to the difference in morphology between the strains in our in vitro experimental conditions. The non-filamentous strain grows almost exclusively as yeast-like cells (Wartenberg et al. 2014), while both oral isolates primarily grow as larger pseudohyphal cells. However, our data with the oral isolates show that our results with the non-filamentous strain can be applied to medically relevant strains in the context of intestinal translocation. These results suggest that disruption to the intestinal barrier due to external factors can foster fungal translocation of C. albicans strains with diminished filamentous growth and damage potential.

We observed that invasive hyphae occasionally reverted back to yeast growth following translocation. The so-called hypha-to-yeast transition has been proposed as yet another mechanism for yeast cells to reach the bloodstream and then further disseminate (Sprague et al. 2022). Our results show that expression of the C. albicans gene PES1 in translocated hyphae is significantly higher than in hyphae only growing atop the IEC barrier. PES1 has previously been shown to be required for disperser cell formation in biofilms and full virulence in systemic candidiasis (Uppuluri et al. 2012, Uppuluri et al. 2018). While the previous in vivo work showed that physiological expression of PES1 is required for full virulence, it used a murine model of infection via tail vein injection, which bypasses the process of intestinal translocation altogether (Uppuluri et al. 2012). Nevertheless, our data show that PES1 expression is heavily associated with translocated hyphae in vitro, where growth of lateral yeast cells from hyphae can be observed. These results suggest that following translocation of C. albicans hyphae across IECs, a morphological transition from hyphal to yeast growth provides another mechanism for yeast cells to reach the bloodstream and disseminate.

Previous studies have shown that while yeast and hyphae persist in ex vivo whole blood models of infection, C. albicans yeast cells form hyphae in human blood (Fradin et al. 2003, Fradin et al. 2005, Machata et al. 2020). Despite this, the transcriptional response is dominated by interactions with neutrophils, which repress the expression of hypha-associated genes (Fradin et al. 2003, Fradin et al. 2005). However, the effect of blood on morphology and gene expression of hyphae has not been systematically investigated. The yeast-to-hypha transition is often a response to new environmental conditions (temperature, pH, nutrients, etc.), so it is likely that the inverse is also true (hypha-to-yeast) (Sudbery 2011). Our in vitro data show that hyphae that have translocated and encounter new conditions, like fresh medium with excess nutrients and no host cells to adhere to, begin expressing PES1 and forming yeast in response. While this is specific to our experimental setup, something similar is to be expected in vivo. Invading hyphae penetrate through the intestinal epithelium, eventually reaching the bloodstream, where they encounter a new environment comprised of different host cells and proteins as well as flow, among many others.

Fungal translocation across the intestinal epithelium requires disruption of the barrier integrity in both in vitro and in vivo models of infection (Koh et al. 2008, Allert et al. 2018, Sprague et al. 2024), though to what degree fungal-mediated damage contributes to this is not clear. Our data demonstrate that C. albicans yeast cells can translocate across the intestinal epithelium together with invading hyphae via a hitchhiking mechanism and passively upon significant loss of barrier integrity. Disruption of the intestinal epithelial barrier increased the translocation efficiency of oral C. albicans isolates that do not readily form hyphae. Invasive hyphae also expressed PES1, a fungal gene essential for yeast growth, following intestinal translocation. This suggests that C. albicans yeast cells can enter the bloodstream from the intestine by hitchhiking with invading hyphae and following a hypha-to-yeast morphological transition.

Conflict of interest

None declared.

Funding

JLS and BH were supported by the German Research Foundation (Deutsche Forschungsgemeinschaft—DFG) within the Collaborative Research Centre (CRC)/Transregio (TRR) 124 “FungiNet” project C1 (project number 210879364). TBS and BH were supported by the DFG under Germany’s Excellence Strategy (EXC 2051–Project-ID 390713860). TL, JS, and BH were supported by the Priority Program SPP2225 “Exit strategies of intracellular pathogens” of the DFG (project number 446404928). SA and BH were supported by the DFG project Hu 528/20-1.

Data availability

The data underlying this article are available in the article and in its online supplementary material.