Candidalysin triggers epithelial cellular stresses that induce necrotic death

Candida albicans is a common opportunistic fungal pathogen that causes a wide range of infections from superficial mucosal to hematogenously disseminated candidiasis. The hyphal form plays an important role in the pathogenic process by invading epithelial cells and causing tissue damage. Notably, the secretion of the hyphal toxin candidalysin is essential for both epithelial cell damage and activation of mucosal immune responses. However, the mechanism of candidalysin‐induced cell death remains unclear. Here, we examined the induction of cell death by candidalysin in oral epithelial cells. Fluorescent imaging using healthy/apoptotic/necrotic cell markers revealed that candidalysin causes a rapid and marked increase in the population of necrotic rather than apoptotic cells in a concentration dependent manner. Activation of a necrosis‐like pathway was confirmed since C. albicans and candidalysin failed to activate caspase‐8 and ‐3, or the cleavage of poly (ADP‐ribose) polymerase. Furthermore, oral epithelial cells treated with candidalysin showed rapid production of reactive oxygen species, disruption of mitochondria activity and mitochondrial membrane potential, ATP depletion and cytochrome c release. Collectively, these data demonstrate that oral epithelial cells respond to the secreted fungal toxin candidalysin by triggering numerous cellular stress responses that induce necrotic death.


| INTRODUCTION
Candida albicans is a human fungal pathogen that causes morbidity and mortality in millions of individuals worldwide each year (Brown et al., 2012). C. albicans possesses a multitude of virulence factors (Richardson, Ho, & Naglik, 2018) but the secretion of the peptide toxin candidalysin (encoded by the ECE1 gene) is a key driver of cell damage and immune responses in mucosal and systemic models of C. albicans infection (Aggor et al., 2020;Allert et al., 2018;Drummond et al., 2019;Ho et al., 2019;Ho et al., 2020;Kasper et al., 2018;Moyes et al., 2016;Swidergall et al., 2019;Verma et al., 2017;Verma et al., 2018). Cell damage often results in cell stress and ultimately death, which can play a key role in host defence to microbial infections (Camilli, Blagojevic, Naglik, & Richardson, 2020;Jorgensen, Rayamajhi, & Miao, 2017). Furthermore, pathogens have evolved several strategies to induce or inhibit host cell death, aiding dissemination and survival within the host. Cell death can be driven by several biological processes including apoptosis, autophagy and different types of necrotic cell death, which can influence the outcome of a microbial insult (Camilli et al., 2020;Jorgensen et al., 2017).
Apoptosis is a highly complex form of programmed cell death, involving an energy-dependent cascade of molecular and cellular events, and can be triggered by intracellular and extracellular stimuli resulting in the activation of intrinsic and extrinsic pathways (Taylor, Cullen, & Martin, 2008). Death receptors, mitochondria and caspases drive an ordered cascade of enzymatic events that promote cell shrinkage, membrane blebbing, nuclear and cytoplasmic condensation, DNA degradation and ultimately fragmentation of the cell into apoptotic bodies, which are rapidly cleared by neighbouring phagocytes (Taylor et al., 2008). In contrast, necrotic cell death is characterised morphologically by cell and organelle swelling, loss of plasma membrane integrity and consequent inflammation. Although necrotic cell death was previously believed to result from injury, and was usually considered to be an uncontrolled process, recent evidence suggests that necrosis can also be tightly regulated (Vanden Berghe, Linkermann, Jouan-Lanhouet, Walczak, & Vandenabeele, 2014).
While induction of apoptosis and necrotic cell death has predominantly been investigated in the context of viral and bacterial infections (Jorgensen et al., 2017), their importance during fungal diseases has only recently been recognised. In addition, data regarding induction or manipulation of cell death pathways during the interplay between pathogenic fungi and host cells are more limited (Camilli et al., 2020).
Thus far, few components have been proposed to play a role in the induction of apoptosis in immune and non-immune cells in response to C. albicans. The cell-wall associated sphingolipid, phospholipomannan, can induce apoptosis in murine macrophages (Ibata-Ombetta, Idziorek, Trinel, Poulain, & Jouault, 2003), while glycan moieties (Wagener et al., 2012) and secreted aspartic proteases (Wu et al., 2013) may trigger apoptosis in oral and lung epithelial cells, respectively. However, while transcript profiling data revealed changes in the expression of apoptotic genes in oral epithelial cells after infection with live C. albicans, only a small fraction (10-15%) of cells exhibited mid-late apoptotic features, including loss of mitochondrial integrity, caspase-9/3 activity and DNA fragmentation (Villar, Chukwuedum Aniemeke, Zhao, & Huynh-Ba, 2012). Moreover, upregulation of anti-apoptotic genes has been observed in both epithelial cells and macrophages in response to C. albicans (Moyes et al., 2014;Reales-Calderon et al., 2013;Villar et al., 2012). Thus, although pro-and anti-apoptotic signalling events have been reported following in vitro stimulation of myeloid and epithelial cells with C. albicans or fungal components, apoptosis is a non-lytic form of cell death and cannot account for C. albicans-induced lysis. Live C. albicans induced early apoptotic signalling events in oral epithelial cells and macrophages followed by necrotic death (Panagio, Felipe, Vidotto, & Gaziri, 2002;Villar & Zhao, 2010). Furthermore, necrotic pathways that promote lytic cell death have been recently described in myeloid cells following Candida infection (Camilli et al., 2020;Cao et al., 2019;Kasper et al., 2018;O'Meara et al., 2015;O'Meara et al., 2018;Tucey et al., 2020;Uwamahoro et al., 2014;Vylkova & Lorenz, 2017;Wellington, Koselny, Sutterwala, & Krysan, 2014).
During infection, C. albicans forms hyphae, which invade tissues and induce cell death. Invasion occurs via two routes, depending on the epithelial cell type: induced endocytosis or active penetration (Dalle et al., 2010;Zakikhany et al., 2007). During the invasion, C. albicans causes epithelial invagination and induces an invasion pocket that is surrounded by the host cell membrane (Zakikhany et al., 2007). It is hypothesised that candidalysin is secreted into the invasion pocket, where it intercalates into host membranes to induce pore formation, which results in cell damage and ultimately, cell death (Naglik, Gaffen, & Hube, 2019). However, whether candidalysin induces epithelial cell death via apoptosis, necrosis or a combination of both mechanisms is currently unclear. Here, we provide evidence that candidalysin induces oxidative stress, mitochondrial dysfunction and epithelial cell death, which occurs predominantly via a necrotic pathway.
2 | RESULTS 2.1 | Candidalysin-induced epithelial cell death does not activate apoptosis-related caspases and occurs through a necrotic mechanism Epithelial cell death involves physical changes to the host cell membrane and disruption of cellular homeostasis. Thus, a qualitative approach was initially used to investigate cell death features in TR146 oral epithelial cells exposed to candidalysin. Epithelial cells were treated with candidalysin at different concentrations and stained after 3, 6 and 24 h with FITC-Annexin V, EthD-III and Hoechst 33342. FITC-Annexin V produces green fluorescence when bound to phosphatidylserine residues exposed on the outer surface of the lipid bilayer of apoptotic cells. EthD-III is unable to cross intact epithelial plasma membranes but can enter cells with damaged membranes, where it intercalates with DNA to produce red/yellow fluorescence. Hoechst 33342 (blue fluorescence) was used to visualise epithelial nuclei. As positive controls, 0.1 μM staurosporine was used to induce apoptosis (Kabir, Lobo, & Zachary, 2002) and 1% (vol/vol) Triton X-100 was used to induce cell necrosis. Fluorescence microscopy revealed markedly higher levels of necrotic cells in the population (EthD-III single positive), when compared to apoptotic cells (FITC-Annexin V single positive) or late apoptotic/ necrotic cells (FITC-Annexin V/EthD-III double positive), following candidalysin stimulation (Figure 1). Plasma membrane damage, as measured by EthD-III incorporation, was observed by 3 h post stimulation with candidalysin, and increased in a concentrationdependent manner at all time points. The data indicate that candidalysin causes cytotoxicity in oral epithelial cells predominantly through necrosis.
To further investigate whether candidalysin-induced cytotoxicity was primarily due to necrosis rather than apoptosis, we analysed the processing and activity of apoptotic caspases.
Caspases are cysteine-aspartic acid proteases which, when cleaved and activated, govern apoptotic cell death. Activation of caspase-8 is an essential step in the initiation of the extrinsic apoptosis pathway, while caspase-3 is the terminal caspase activated by both extrinsic and intrinsic pathways. We, therefore, examined the activation of caspase-8 and -3 in candidalysin-mediated cell death. Only epithelial cells treated with staurosporine exhibited a modest cleavage of immature caspase-8 (57 kDa) or caspase-3 (31 kDa) to produce products consistent with the molecular weight of active caspase-8 (17 kDa) and caspase-3 (19/17 kDa) (white arrows; Figure 2A, B). Since pro-caspase-8 and procaspase-3 were strongly expressed in epithelial cells but only modestly activated following staurosporine treatment, we also used a more sensitive luminescence-based assay to detect the presence of active caspases. Neither C. albicans strains nor candidalysin induced caspase-8 or caspase-3 activity in epithelial cells when compared with the vehicle control ( Figure 2c). In contrast, staurosporine induced significant caspase-8 and caspase-3 F I G U R E 1 Cytotoxicity profile of candidalysin-treated epithelial cells. Fluorescence microscopy images of TR146 oral epithelial cells exposed to the indicated concentrations of candidalysin for 3, 6 and 24 h. Staurosporine and Triton X-100 were used as positive controls to induce apoptosis and necrosis, respectively. Cells were stained with Hoechst 33342 (blue: live cells), FITC-conjugated Annexin V (green: apoptotic cells) and EthD-III (red/yellow: necrotic cells). Data are representative of 3 independent experiments. Images were taken with a fluorescence microscope at Â100 magnification. Scale bars = 500 μm activity ( Figure 2c). Collectively, these data demonstrate that while TR146 oral epithelial cells express high baseline levels of pro-caspase-8 and pro-caspase-3, neither C. albicans nor candidalysin induce the production of active caspase-8 and caspase-3 from their respective pro-caspases.
During apoptosis, poly (ADP-ribose) polymerase (PARP) is inactivated by caspase-3-mediated cleavage, which prevents DNA repair within the cell (Kaufmann & Kaufman, 1993). To confirm that a caspase-3-independent cell death mechanism was occurring in response to C. albicans and candidalysin, we examined the cleavage of PARP by western blot. In contrast to staurosporine, treatment of oral epithelial cells with C. albicans or candidalysin was not observed to induce the PARP apoptotic signature characterised by the appearance of an 89 kDa fragment (Figure 2d). Taken together, these data confirm that apoptotic events are not features of the epithelial response to C. albicans and candidalysin.

| Candidalysin disrupts epithelial metabolic activity and induces reactive oxygen species production and depletion of intracellular ATP
We next sought to identify the subcellular events in candidalysin-induced cell death in oral epithelial cells. We previously reported that candidalysin causes epithelial cell damage and induces calcium influx (Ho et al., 2019;Moyes et al., 2016), which are characteristics associated with cell stress and death (Cerella, Diederich, & Ghibelli, 2010). Interestingly, elevation of intracellular calcium in response to candidalysin appears to depend predominantly on calcium influx from the extracellular environment, since the toxin failed to induce an increase in intracellular calcium levels in calcium-free buffer ( Figure S1). Furthermore, alterations in cellular metabolism, adenosine triphosphate (ATP) and reactive oxygen species (ROS) are common features of cytotoxicity in several cell death pathways, including necrotic death (Zong & Thompson, 2006). Accordingly, we Collectively, these data indicate that candidalysin induces a loss of metabolic activity, depletion of ATP and oxidative stress in epithelial cells which is closely associated with a necrotic cell death mechanism.

| Candidalysin causes mitochondrial dysfunction and cytochrome c release during C. albicans infection
Mitochondria play a pivotal role in the production of energy (e.g., ATP) and mitochondrial dysfunction is a central feature of both apoptotic and necrotic cell death (Baines, 2010;Wang, 2001). Therefore, we investigated whether loss of mitochondrial membrane potential (ΔΨm) and/or loss of mitochondrial membrane integrity occurs in epithelial cells in response to candidalysin. We used rhodamine 123 (a green fluorescent dye that selectively accumulates in mitochondria in a membrane potential-dependent manner) and confocal microscopy to visualise changes in mitochondrial potential induced by C. albicans and candidalysin. A concentration of 30 μM candidalysin was chosen to minimise membrane permeability, cell swelling and detachment, rapidly and robustly induced by higher concentration of the toxin (e.g., 70 μM).
Epithelial cells treated with 30 μM candidalysin exhibited a reduction in ΔΨm compared with untreated control cells (Figure 4a). Since candidalysin induces robust calcium influx (Ho et al., 2019;Moyes et al., 2016), and calcium levels affect mitochondrial dynamics (Finkel et al., 2015), we also conducted experiments in a calcium free buffer.
Notably, candidalysin-induced mitochondria depolarization was inhibited in the absence of calcium, strongly suggesting that the toxin induces changes in mitochondrial membrane potential in response to increasing levels of cytosolic calcium (Figure 4a). Similarly, we found that mitochondria depolarization was markedly higher following infection of oral epithelial cells with wild type C. albicans as compared with a candidalysin deficient C. albicans strain (ece1Δ/Δ; Figure 4b).
Interestingly, it appears that candidalysin mediates its effects on mitochondria from the epithelial surface, as an AlexaFluor 647-labelled candidalysin (also damaging) remained in the plasma membrane and did not enter the epithelial cell ( Figure 4c). Taken together, these data demonstrate that during epithelial infection, candidalysin induces significant cell stress by affecting mitochondrial dynamics in a calciumdependent manner.
Dysfunctional mitochondria release proteins including cytochrome c that amplify intracellular caspase cascades during apoptosis.
However, release of cytochrome c into the cytosol also occurs during necrotic cell death (Li, Li, Pinto, & Pardee, 1999). Moreover, the rapid disappearance of cytochrome c from dying cells has been attributed to its release into the extracellular environment (Jemmerson, LaPlante, & Treeful, 2002). Therefore, although we observed no caspase activation in candidalysin-mediated cell death, the loss of mitochondrial membrane potential following candidalysin treatment prompted us to investigate whether cytochrome c is released from epithelial cells.
Epithelial cells were treated with candidalysin or C. albicans C. albicans is known to induce both necrotic and apoptotic cell death mechanisms in human cells (Camilli et al., 2020), the role of candidalysin in driving these responses was unclear. Using an in vitro model of intestinal translocation, C. albicans was observed to associate with fungus-induced necrotic, but not apoptotic, intestinal epithelial cell death (Allert et al., 2018). Importantly, candidalysin was observed to cause enterocyte damage, which correlates with fungal translocation (Allert et al., 2018). Here, we demonstrate that C. albicans/candidalysin damages human oral epithelial cells in a manner that is independent of apoptotic caspases, but which exhibits several hallmarks of necrotic death. Furthermore, we provide new insight into the mechanisms by which candidalysin may orchestrate a necrotic response in epithelial cells.
Multiple cellular assays demonstrated that lytic concentrations of candidalysin induce a necrotic effect and appears to target mitochondrial function, as evidenced by decreased metabolic activity, depletion of intracellular ATP and an increase in the level of intracellular ROS. The effect on mitochondria was further confirmed by confocal microscopy which revealed destabilisation of mitochondrial membrane potential when epithelial cells were exposed to wild type C. albicans or candidalysin, but not with a C. albicans strain unable to produce candidalysin. Notably, candidalysin was not observed to enter epithelial cells during the 10 min analysis period, suggesting that the toxin mediates its effects from the cell surface. Furthermore, these cellular effects were only observed in the presence of calcium. Given that calcium influx is a key feature of can-  (Camilli et al., 2020). Early apoptotic events have been described in macrophages in response to the highly virulent CR1 strain of C. albicans in vitro and in vivo (Gasparoto, Gaziri, Burger, de Almeida, & Felipe, 2004;Panagio et al., 2002). However, phagocytosis of CR1 yeast cells induced early apoptotic events before germ tube formation and subsequent lysis of macrophages by necrosis (Panagio et al., 2002). Consistent with the findings from macrophages, live C. albicans induced early apoptotic signalling events in oral epithelial cells followed by necrotic death (Villar & Zhao, 2010). Only a modest increase of caspase-3 and -9 (but not caspase-8) activity was observed during the early stages of infection, which returned to levels similar to those observed in uninfected cells at later time points (Villar & Zhao, 2010). The morphological transition from yeast to hyphae is, therefore, likely to play a crucial role in the inhibition of apoptotic progression and induction of a lytic cell death mechanism.
Accordingly, our data indicate that oral epithelial cells do not undergo pro-apoptotic changes in response to the hyphal toxin candidalysin.
Our study describes the responses of a human buccal squamous cell carcinoma cell line to C. albicans and candidalysin. Although one of the hallmarks of cancer cells is resistance to apoptosis, several studies have demonstrated that TR146 cells are capable of undergoing apoptosis (Han et al., 2015;O'Callaghan et al., 2015;Wagener et al., 2012). Similarly, in the present study, we demonstrate that staurosporine induces annexin V binding, caspase-3 and -8 activation and cleavage of PARP in TR146 cells. However, the cells were most intensively stained with EthD-III following candidalysin treatment, and candidalysin did not induce caspase-3 or -8 processing, or detectable caspase activity at any concentration tested. These data suggest that cell death proceeds through a lytic but non-apoptotic mechanism.
Recent research with vaginal epithelial cells further corroborate these findings (Pekmezovic et al., 2021). Staurosporine was observed to trigger apoptosis in both primary vaginal cells and a vaginal epithelial carcinoma cell line, while the level of apoptosis was not observed to differ between Candida infected and uninfected cells (Pekmezovic et al., 2021). Moreover, primary mononuclear phagocytes undergo necrotic cell death in response to candidalysin with only minimal exposure of cell surface phosphatidylserine (Annexin V-positive cells) in the absence of pro-apoptotic caspase activation (Kasper et al., 2018). A similar mechanism may, therefore, contribute to candidalysin-induced cell death in both myeloid and epithelial cells. inflammasome-dependent pyroptosis (Camilli et al., 2020;O'Meara et al., 2015O'Meara et al., , 2018Tucey et al., 2020;Uwamahoro et al., 2014;Wellington et al., 2014). However, although candidalysin activates the NLRP3 inflammasome and induces cytolysis in both murine and human mononuclear phagocytes, cell death appears to be independent of inflammasome activation, as demonstrated by the use of a caspase-1 inhibitor or mononuclear phagocytes isolated from mice lacking NLRP3 or caspase-1 (Kasper et al., 2018). Furthermore, the necroptosis inhibitor necrostatin-1 did not inhibit the cytolytic effect of candidalysin (Kasper et al., 2018), suggesting that neither necroptosis nor pyroptosis contribute to candidalysin-driven macrophage cell death. Whether C. albicans/candidalysin kills macrophages through a different regulated necrotic pathway and whether the cell death mechanism is conserved between myeloid and non-myeloid cells remains to be determined.
It is now widely accepted that mitochondria represent a central control point not only during apoptosis but also for the execution of necrotic cell death programmes (Baines, 2010). Numerous studies have suggested that calcium overload and oxidative stress lead to the opening of a channel in the mitochondria inner membrane, termed the mitochondrial permeability transition pore (mPTP), which results in a dramatic depolarization of ΔΨm and release of proteins and solutes including apoptogenic factors (Bauer & Murphy, 2020). Notably, a sustained and prolonged mPTP opening is associated with the loss of mitochondrial membrane potential, cessation of ATP synthesis, mitochondrial swelling, rupture and necrotic cell death (Ying & Padanilam, 2016). Indeed, while apoptosis is an ATP-dependent process, necrosis is ATP-independent and therefore, a high proportion of damaged mitochondria and ATP depletion are likely to promote necrotic, but not apoptotic cell death (Ying & Padanilam, 2016). Similarly, our data establish that candidalysin secretion by C. albicans is the key trigger of cell death in epithelial cells, a process that is closely linked to calcium influx, oxidative stress, mitochondrial dysfunction and decreased metabolic activity. As such, the implications of mitochondrial dysfunction in candidalysin-induced cell death may be of particular interest for future investigations.
In summary, this study demonstrates that candidalysin induces a specific pro-death pathway from the cell surface that converges on mitochondria to regulate a necrotic cell death program in epithelial cells. Further studies will be necessary to identify the precise molecular pathways that promote the alteration of mitochondrial physiology in epithelial cells during C. albicans/candidalysin infection.

| Cell culture
The TR146 human buccal epithelial squamous cell carcinoma cell line (Rupniak et al., 1985) was obtained from the European Collection of Authenticated Cell Cultures (ECACC). TR146 cells were cultured in Dulbecco's Modified Eagle Medium/F-12 (DMEM/F12) Nutrient Mixture (1:1) + L-glutamine (Life Technologies) supplemented with 15% (vol/vol) heat-inactivated foetal bovine serum (Life Technologies) and 1% (vol/vol) penicillin-streptomycin (Sigma) at 37 C, 5% CO 2 . For cell damage, enzyme linked enzyme-linked immunoassay (ELISA) and western blotting assays, the culture medium was replaced 24 h prior to experimentation with serum-free medium and maintained until the cells were harvested.

| C. albicans strains
All C. albicans strains used in this study were generated previously (Moyes et al., 2016) and are listed in Table 1 and magnesium, and medium replaced with isotonic minimal media containing calcium. Log-phase C. albicans was added to monolayers at an MOI of 0.25; the plates were then centrifuged for 1 min at 1500 rpm and incubated for 4 h at 37 C. Following invasion, coverslips were stained for 5 min with 10 μg/ml calcofluor white to label C. albicans and coverslips were viewed live.
4.10 | Imaging of candidalysin-Alexa647 treated cells • One day prior to experiments, TR146 cells were seeded onto 18 mm glass coverslips at a concentration of 1 x 10 5 cells/ml in DMEM/F12 supplemented with L-glutamine and 10% heat-inactivated foetal calf serum, and incubated 16 h at 37 C, 5% CO 2 . The following day, cells were washed with 1X PBS and incubated in isotonic minimal media containing calcium. Synthetic candidalysin-Alexa647, previously resuspended to a concentration of 1.4 mM in sterile dH 2 0, was cup sonicated in ice water for 5-60 s intervals to obtain a homogeneous suspension. TR146 monolayers were then incubated with 30 μM candidalysin-Alexa647 for 10 min at 37 C. After incubation, cells were washed three times with 1X PBS and fixed in 4% paraformaldehyde for 10 min at room temperature. Fixative was replaced with 1X PBS and coverslips viewed by confocal microscopy.

| Confocal microscopy and image analysis
Confocal images were acquired using a Yokogawa CSU10 spinning disk system (Quorum Technologies, Inc.

| Detection of caspase activity
Caspase-3 and Caspase-8 activity was measured using a Caspase-Glo 3/7 and Caspase-Glo 8 Assay (Promega), according to the manufacturer's protocol. Briefly, a solution containing a cell lysis reagent and a quenched (non-luminescent) caspase-specific substrate was added to control and treated cells (1:1 ratio of assay reagent volume to sample volume). Luminescence was measured after a 30 min incubation with substrate for caspase-3 and 1 h incubation for caspase-8 using a FlexStation 3 (Molecular Devices). Data was analysed using Softmax Pro software.

| Statistical analysis
The significance level was determined by one-way analysis of variance (ANOVA) with Dunnett's correction for multiple comparisons. A difference was considered statistically significant at p < .05. At least three independent biological replicate experiments were performed.
Data were analysed using GraphPad Prism, version 8.0 (GraphPad Software) and are shown as the mean ± SEM.