Oxysterols Protect Epithelial Cells Against Pore-Forming Toxins

Many species of bacteria produce toxins such as cholesterol-dependent cytolysins that form pores in cell membranes. Membrane pores facilitate infection by releasing nutrients, delivering virulence factors, and causing lytic cell damage - cytolysis. Oxysterols are oxidized forms of cholesterol that regulate cellular cholesterol and alter immune responses to bacteria. Whether oxysterols also influence the protection of cells against pore-forming toxins is unresolved. Here we tested the hypothesis that oxysterols stimulate the intrinsic protection of epithelial cells against damage caused by cholesterol-dependent cytolysins. We treated epithelial cells with oxysterols and then challenged them with the cholesterol-dependent cytolysin, pyolysin. Treating HeLa cells with 27-hydroxycholesterol, 25-hydroxycholesterol, 7α-hydroxycholesterol, or 7β-hydroxycholesterol reduced pyolysin-induced leakage of lactate dehydrogenase and reduced pyolysin-induced cytolysis. Specifically, treatment with 10 ng/ml 27-hydroxycholesterol for 24 h reduced pyolysin-induced lactate dehydrogenase leakage by 88%, and reduced cytolysis from 74% to 1%. Treating HeLa cells with 27-hydroxycholesterol also reduced pyolysin-induced leakage of potassium ions, prevented mitogen-activated protein kinase cell stress responses, and limited alterations in the cytoskeleton. Furthermore, 27-hydroxycholesterol reduced pyolysin-induced damage in lung and liver epithelial cells, and protected against the cytolysins streptolysin O and Staphylococcus aureus α-hemolysin. Although oxysterols regulate cellular cholesterol by activating liver X receptors, cytoprotection did not depend on liver X receptors or changes in total cellular cholesterol. However, oxysterol cytoprotection was partially dependent on acyl-CoA:cholesterol acyltransferase (ACAT) reducing accessible cholesterol in cell membranes. Collectively, these findings imply that oxysterols stimulate the intrinsic protection of epithelial cells against pore-forming toxins and may help protect tissues against pathogenic bacteria.


INTRODUCTION
Many species of pathogenic bacteria produce toxins that form pores in the cell membrane of eukaryotic cells (1,2). These pores facilitate infection by releasing nutrients, delivering virulence factors, and causing lytic cell damage -cytolysis. The most common pore-forming toxins are cholesterol-dependent cytolysins, which bind to cholesterol in cell membranes (2)(3)(4)(5). Oxysterols are oxidized forms of cholesterol that help regulate cellular cholesterol (6,7). In addition, some oxysterols regulate immune responses to bacteria and influence disease pathogenesis (8)(9)(10). However, an unresolved question is whether oxysterols can also alter the intrinsic protection of epithelial cells against cholesterol-dependent cytolysins.
Here we tested the hypothesis that oxysterols stimulate the intrinsic protection of epithelial cells against damage caused by cholesterol-dependent cytolysins. We initially treated HeLa cells with oxysterols and then subsequently challenged the cells with pyolysin. Protection of cells against damage caused by pyolysin was assessed by measuring the leakage of potassium and LDH, and by evaluating MAPK stress responses, cell morphology and cell viability. Evidence for oxysterol cytoprotection was also examined for other types of epithelial cells and pore-forming toxins. In addition, we investigated whether the increased protection of cells against pyolysin following treatment with oxysterols was mechanistically associated with ion flux, LXR expression, or ACAT activity.

Pore-Forming Toxins
The plo plasmid (pGS59) was a gift from Dr H Jost (University of Arizona), and the pyolysin protein was generated and purified as described previously, with a specific activity of 628,338 hemolytic units (HU)/mg protein (21,30). Streptolysin O was purchased, stored as a 1 mg/ml solution, and activated with 10 mM dithiothreitol according to the manufacturer's instructions (Merck). Staphylococcus aureus a-hemolysin was purchased and stored as a 0.5 mg/ml solution, according to the manufacturer's instructions (Merck).

Cytoprotection
To determine the amount of pyolysin required to study cytoprotection, cells were cultured in serum-free medium for 24 h and then challenged for 2 h with control serum-free medium or medium containing the range of concentrations of pyolysin specified in Results. After the challenge, we measured the leakage of LDH into cell supernatants and assessed cytolysis by measuring viable cells using an MTT assay. For subsequent cytoprotection experiments we selected pyolysin challenges that increased LDH leakage and caused cytolysis in HeLa (100 HU/ well), A549 (25 HU/well), Hep-G2 (100 HU/well) and NC1-H441 cells (200 HU/well).
To examine whether oxysterols protected against pyolysin, cells were treated in serum-free medium with vehicle (0.1% methanol), 27-hydroxycholesterol, 25-hydroxycholesterol, 7ahydroxycholesterol, or 7b-hydroxycholesterol (Avanti Polar Lipids, Alabama, United States), using the concentrations and durations specified in Results. After the treatment period, supernatants were discarded, and cells were challenged for 2 h, with control serum-free medium or pyolysin, in the absence of further treatment. To determine whether cytoprotection extended beyond pyolysin, HeLa and A549 cells were treated for 24 h with vehicle or 27-hydroxycholesterol, and then challenged with control serum-free medium, or streptolysin O or a−hemolysin as specified in Results. The concentration and duration of a-hemolysin challenge was determined by culturing cells for 24 h in serum-free medium and then challenging them for 0 to 24 h using the concentrations specified in Results, which were informed by a−hemolysin concentrations used in HeLa cells previously (31). To determine if pyolysin binds directly to oxysterols, pyolysin was mixed with vehicle, 10 ng/ml 27hydroxycholesterol, or 10 ng/ml 25-hydroxycholesterol, and then used to challenge cells for 2 h. At the end of the cytolysin challenge period, supernatants were collected to measure LDH leakage and viable cells were evaluated by MTT assay or immunofluorescent microscopy.

MAPK Cell Stress Responses
To explore whether oxysterols could protect against pyolysininduced MAPK activation, 1.5 × 10 5 HeLa cells/well were cultured in complete medium for 24 h using 6-well plates, and treated in serum-free medium for 24 h with vehicle, 10 ng/ml 27hydroxycholesterol, or 50 nM T0901317 (Tocris, Abingdon, UK), which is an LXR agonist (32). Cells were then challenge with control serum-free medium or 100 HU pyolysin for 10 min, the supernatants discarded, and the cells washed with 300 ml ice cold PBS and lysed with 100 ml PhosphoSafe Extraction Reagent (Novagen, Darmstadt, Germany) for Western blotting of the MAPK pathway.

Cytoprotection and Ion Flux
To examine the role of potassium efflux or calcium influx on oxysterol protection against pyolysin, HeLa cells were cultured in serum-free medium with vehicle or 10 ng/ml 27hydroxycholesterol for 24 h. Then, for potassium experiments, pyolysin was prepared in low-potassium medium ( to prevent potassium efflux, as described previously (33). Cells were washed twice with low-potassium or high-potassium medium prior to a 2 h challenge with pyolysin in the cognate medium. For calcium experiments, the cells were washed twice with calcium-free Dulbecco's phosphate-buffered saline (DPBS, Thermo Fisher Scientific). Cells were then challenged for 2 h with or without pyolysin in control medium (1.8 mM CaCl) or calcium-free medium (Thermo Fisher Scientific) to prevent calcium influx. At the end of the challenge period, supernatants were collected to measure LDH leakage and viable cells were evaluated by MTT assay.

Cytoprotection and Liver X Receptors
To examine the role of LXRa and LXRb in cytoprotection, we used siRNA to target NR1H3 and NR1H2, prior to treatment with 27-hydroxycholesterol and a subsequent pyolysin challenge. We used the LXR agonist T0901317 as a positive control, with the concentration determined by treating cells with the range of concentrations specified in Results and measuring LDH leakage and cell viability following a 2 h pyolysin challenge. Briefly, HeLa cells were transfected with scramble siRNA (ON-TARGETplus Non-targeting Control Pool; Horizon discovery, Cambridge, UK), or siRNA designed using Dharmacon siDesign Centre to target NR1H3 and NR1H2 (Supplementary Table 1). A mixture of 20 pmol of siRNA, 100 ml Opti-MEM 1 medium and 1.5 µl Lipofectamine RNAiMAX Reagent (both Thermo Fisher Scientific) were added to each well of a 24-well plate and incubated for 20 min, and then 4 × 10 4 HeLa cells/well were seeded in 900 µl DMEM medium supplemented with 10% FBS for 48 h. Supernatants were discarded and cells were treated with vehicle, 10 ng/ml 27-hydroxycholesterol or 50 nM T0901317 for 24 h in serum-free culture medium prior to a 2 h challenge with control medium or pyolysin. At the end of the challenge period, supernatants were collected to measure LDH leakage and cell viability was evaluated with MTT assay.

Cytoprotection and Cholesterol
To investigate the effect of oxysterol treatment on pyolysin binding to accessible cholesterol in the cell membrane, 1.5 × 10 5 HeLa cells/well were cultured in complete medium for 24 h using 6-well plates, and then treated in serum-free medium with vehicle, 10 ng/ml 27-hydroxycholesterol or 50 nM T0901317 for 24 h, before a 2 h challenge with control serum-free medium or pyolysin. The cells were washed with 300 ml ice cold PBS and lysed with 100 ml PhosphoSafe Extraction Reagent for Western blotting for pyolysin binding.
To examine the role of serum cholesterol in cytoprotection against pyolysin we first cultured HeLa cells for 24 h in medium containing 10% serum with vehicle or 25-hydroxycholesterol, followed by challenge with control medium or pyolysin for 2 h. To verify protection against pyolysin by statins inhibiting HMGCR (34), HeLa cells were cultured for 24 h in serum-free medium with vehicle or 10 µM atorvastatin (Merck), followed by challenge with control medium or pyolysin for 2 h. To examine the role of ACAT esterification of cholesterol in cytoprotection against pyolysin, we used the selective ACAT inhibitor Sandoz 58-035 (SZ58-035, Merck), as described previously (10,35). Cells were washed twice with phosphate-buffered saline (PBS, Thermo Fisher Scientific), and cultured in serum-free medium with vehicle (0.1% DMSO) or 10 µM SZ58-035 for 16 h. The cells were then washed twice with PBS, and treated with vehicle, 10 ng/ml 27-hydroxycholesterol or 10 ng/ml 25−hydroxycholesterol for 24 h in the medium containing DMSO or SZ58-035, followed by challenge with control medium or pyolysin, for 2 h. At the end of the challenge periods, supernatants were collected to measure LDH leakage, and cell viability was measured using an MTT assay.
To explore the effect of increasing membrane accessible cholesterol on protection against pyolysin, sphingomyelincholesterol complexes were disrupted using sphingomyelinase from Staphylococcus aureus (Merck), as described previously (5,10). Cells were treated in serum-free medium with vehicle (0.1% methanol) or 10 ng/ml 27-hydroxycholesterol for 24 h. The cells were washed twice with PBS and cultured in serum-free medium with or without 100 mU/ml sphingomyelinase for 30 min, washed twice with PBS, and then challenged with control medium or 25 HU pyolysin for 2 h. In preliminary experiments, we titrated the amount of pyolysin used to challenge the cells because the cells were more susceptible to pyolysin following sphingomyelinase treatment, with 100 HU pyolysin causing excessive cell lysis (> 97%), as determined by MTT assay. At the end of the challenge periods, supernatants were collected to measure LDH leakage, and cell viability was measured using an MTT assay.

Lactate Dehydrogenase
Cell supernatant LDH was quantified by LDH−dependent conversion of lactate to pyruvate, via reduction of b−Nicotinamide adenine dinucleotide sodium salt (NAD+) to NADH, which is detected by NADH−dependent reduction of a tetrazolium salt to formazan, as described previously (18). Inter− and intra−assay coefficients of variation were < 4%.

Potassium
The leakage of potassium was measured as described previously (17,29). Briefly, 1.5 × 10 5 cells/well were cultured for 24 h in complete medium using 6-well culture plates, and then treated with 27-hydroxycholesterol or 25-hydroxycholesterol for 24 h using the concentrations specified in Results. The cells were washed and challenged in potassium-free buffer for 5 min with the amounts of pyolysin or streptolysin O indicated in Results, or for 15 min with 8 µg a-hemolysin (no potassium leakage was detected after 5 min of a-hemolysin challenge). Extracellular potassium was measured in cell supernatants using a Jenway PFP7 flame photometer (Cole-Parmer, Stone, Staffordshire, UK). The inter-and intra-assay coefficients of variation were < 4%.

Cholesterol
To measure cellular cholesterol, 1.5 × 10 5 HeLa cells/well were cultured in complete medium for 24 h using 6-well culture plates (TPP), and then treated in serum-free medium with 10 ng/ml 27hydroxycholesterol, 10 ng/ml 25-hydroxycholesterol, 1 mM methyl-b-cyclodextrin (Merck) or 10 µM atorvastatin for 24 h. The cells were then washed twice with PBS, collected in 200 µl/ well cholesterol assay buffer (Thermo Fisher Scientific), and stored at -20⁰C. Cellular cholesterol was measured using the Amplex Red Cholesterol Assay Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. Cholesterol concentrations were normalized to total protein concentrations, measured using a DC protein assay (Bio-Rad Laboratories, Hercules, CA, United States), as described previously (18,29). The inter-and intra-assay coefficients of variation for the cholesterol assay were < 5% and < 6% respectively.

Immunofluorescent Microscopy
We stained cells using phalloidin to visualize cytoskeletal and cell damage, as described previously (17). Briefly, 4 × 10 4 cells/well were cultured on glass cover slips in complete-medium for 24 h using 24-well culture plates, and then treated for 24 h in serumfree medium with vehicle, 27-hydroxycholesterol, 25hydroxycholesterol or T0901317 as specified in Results. The cells were then challenged with control medium or pyolysin for 2 h, or a-hemolysin for 24 h. At the end of the challenge period, cells were washed with PBS, fixed with 4% paraformaldehyde (Merck), washed with PBS, and permeabilized in 0.2% Triton-X 100. The cells were then blocked for 30 min in PBS containing 0.5% bovine serum albumin (BSA; Merck) and 0.1% Triton-X 100, and then incubated for 1 h with Alexa Fluor 555-conjugated phalloidin (Thermo Fisher Scientific). Cells were washed with 0.1% Triton-X 100 in PBS three times and mounted onto microscope slides using 40,6 diamidino-2-phenylindole to visualize cell nuclei (Vectashield with DAPI; Vector Laboratories Inc., Burlington, CA, USA). The cells were examined using an Axio Imager M1 fluorescence microscope and images captured using an AxioCamMR3 (Zeiss, Jena, Germany). The proportion of cells that had cytoskeletal changes (cytoskeletal contraction, disrupted shape, or loss of actin fiber definition) were counted using > 135 cells per treatment across 3 independent images per replicate.

Quantitative PCR
Total RNA was isolated from HeLa cells lysed in RLT buffer using the RNeasy Mini Kit according to the manufacturer's instructions (QIAGEN, Crawley, UK). The RNA was quantified with a Nanodrop spectrophotometer (Labtech, Ringmer, UK), and reverse transcription of 1 µg mRNA was performed using the Quantitect Reverse Transcription Kit according to the manufacturer's instructions (QIAGEN). Primers were designed using the NCBI primer design tool for NR1H2, NR1H3 and the reference gene ribosome-like protein 19 (RPL19; Supplementary Table 2). Quantitative PCR was performed in 25 µl reaction volumes comprising 12.5 µl QuantiFast SYBR Green PCR Master Mix (QIAGEN), 0.25 µL forward primer, 0.25 µl reverse primer, 10.5 µl nuclease-free water, and 1.5 ml cDNA on white low-profile 96-well plates (Bio-Rad Laboratories). Thermal cycling parameters were a cycle of 95°C for 5 min, followed by 40 cycles of 95°C for 10 s and 60°C for 30 s, using the CFX Connect Real-time thermal cycler (Bio-Rad Laboratories). Gene expression analysis was performed according to the MIQE guidelines (36). Standard curves were generated from serial dilutions of mRNA extracted from untreated HeLa cells. Gene expression was analyzed in triplicate using the appropriate standard curve quantification cycle, and mRNA expression normalized to RPL19.

Statistical Analysis
The statistical unit was each independent passage of cells. Statistical analysis was performed using GraphPad Prism 9.0.1 (GraphPad Software, San Diego, California, USA), with significance attributed when P < 0.05, and data presented as arithmetic mean ± s.e.m. Comparisons between treatments were made using independent ttest, or using ANOVA followed by Dunnett, Bonferroni or Tukey post hoc test for multiple comparisons, as specified in Results.

Oxysterols Protect HeLa Cells Against Pyolysin
We first determined the amount of pyolysin required to damage HeLa cells. We used HeLa cells because these epithelial cells are often used to explore responses to cholesterol-dependent cytolysins (14,18,24,33). Although culturing cells in medium containing 10% serum did not significantly alter pyolysin-induced leakage of LDH or cytolysis compared with serum-free medium (Supplementary Figure 1), we used serum-free medium to avoid pyolysin binding to serum cholesterol, or serum altering cholesterol homeostasis or inducing MAPK phosphorylation (39,40). We used a 2 h challenge because we aimed to examine whether oxysterols stimulated the intrinsic protection of cells against damage caused by cytolysins, whereas a longer challenge might also reflect membrane repair and cell replication (16,21,30). Pyolysin caused pore formation as determined by the leakage of LDH into supernatants from HeLa cells cultured in 24-well plates, and caused cytolysis as determined by reduced cell viability (Supplementary Figure 2A). For subsequent cytoprotection experiments, we used 100 HU pyolysin per well because this reduced HeLa cell viability by > 80%.
We next investigated whether treatment with the side-chain hydroxycholesterols 27-hydroxycholesterol or 25-hydroxycholesterol protected HeLa cells against a subsequent pyolysin challenge. The cells were treated for 24 h with a range of oxysterol concentrations, informed by human plasma concentrations (41), and then challenged with pyolysin for 2 h without further oxysterol treatment. Both 27hydroxycholesterol and 25-hydroxycholesterol reduced pyolysininduced leakage of LDH and cytolysis ( Figures 1A, B). Specifically, treatment with 10 ng/ml 27-hydroxycholesterol for 24 h reduced LDH leakage by 88% when cells were subsequently challenged with pyolysin for 2 h, and reduced cytolysis from 74% to 1%, whilst 10 ng/ ml 25-hydroxycholesterol reduced pyolysin-induced LDH leakage by 89% and cytolysis from 68% to 2%.
We also examined whether treatment with ring-modified hydroxycholesterols for 24 h might protect HeLa cells against a subsequent 2 h pyolysin challenge. Both 7a-hydroxycholesterol and 7b-hydroxycholesterol reduced pyolysin-induced LDH leakage and cytolysis, but neither were as effective as an equivalent concentration of 27-hydroxycholesterol ( Figures 1F,  G). Collectively these data provide evidence that oxysterols protect HeLa cells against damage caused by pyolysin. As sidechain hydroxycholesterols were most protective, and 27hydroxycholesterol is abundant in human plasma (41), we focused on their ability to stimulate the intrinsic protection of epithelial cells against cytolysins.

27-Hydroxycholesterol Prevents MAPK Phosphorylation
Pyolysin induces a MAPK cell stress response, including pyolysin-induced phosphorylation of ERK, JNK and p38 in HeLa cells (18,24). Treatment with 27-hydroxycholesterol reduced pyolysin-induced ERK 1/2, p38 and JNK phosphorylation ( Figures 3A, B and Supplementary Figure 3). The reduction in pyolysin-induced MAPK phosphorylation with 27hydroxycholesterol was as effective as treatment with an LXR agonist (T0901317), which was used as a positive control because LXR agonists protect bovine endometrial cells and murine macrophages against cholesterol-dependent cytolysins (28,29). These data provide evidence that 27-hydroxycholesterol prevents activation of MAPK cell stress responses to pyolysin.

27-Hydroxycholesterol Protects Cells Against Streptolysin O and a-Hemolysin
To explore whether 27-hydroxycholesterol cytoprotection extended beyond pyolysin we used Streptococcus pyogenes streptolysin O and Staphylococcus aureus a-hemolysin. Streptolysin O is a cholesterol-dependent cytolysin that forms 26 nm diameter b-barrel transmembrane pores, damages cells and causes cytolysis (44,45). Challenging HeLa cells with Streptolysin O for 2 h caused pore formation and cytolysis as expected, but treatment with 25 ng/ml 27-hydrocycholesterol for 24 h reduced streptolysin-induced leakage of LDH, cytolysis, and leakage of potassium ( Figures 4A, B).

Potassium and Calcium Do Not Affect 27-Hydroxycholesterol Cytoprotection
We next investigated potential mechanisms of side-chain hydroxycholesterol cytoprotection against pyolysin. Oxysterols have pleiotropic effects, including altering ion fluxes across cell membranes, increasing MAPK phosphorylation, activating LXRs, accelerating HMGCR degradation, and stimulating ACAT cholesterol esterification (7,10,48,49). We first considered whether oxysterol cytoprotection was dependent on potassium or calcium ion fluxes because these are important for activating cell repair mechanisms (13,16). Challenging HeLa cells with pyolysin in 140 mM high-potassium medium to prevent potassium efflux resulted in reduced LDH leakage and cytolysis compared with challenging cells in a 5 mM lowpotassium medium, but the cytoprotective effect of 27hydroxycholesterol was unaffected ( Figure 5A). Challenging HeLa cells with pyolysin in calcium-free medium to prevent calcium influx resulted in increased LDH leakage and cytolysis, but the cytoprotective effect of 27-hydroxycholesterol was not significantly affected ( Figure 5B).

Liver X Receptors Are Not Required for 27-Hydroxycholesterol Cytoprotection
Oxysterols such as 27-hydroxycholesterol activate LXRa and LXRb, encoded by NR1H3 and NR1H2, respectively (50). To test whether 27-hydroxycholesterol cytoprotection was dependent on LXRs, we transfected HeLa cells with siRNA targeting NR1H3 and NR1H2, which reduced expression by > 80% ( Figure 6A). However, siRNA targeting NR1H3, NR1H2, or NR1H3 and NR1H2 in combination, did not significantly diminish 27-hydroxycholesterol cytoprotection against pyolysin, compared with cells treated with scramble siRNA ( Figure 6B). To verify that the siRNA was effective we used the LXRa and LXRb agonist T0901317 (32). We first established that treating HeLa cells with 50 nM T0901317 for 24 h reduced pyolysin-induced leakage of LDH and cytolysis ( Figure 6C). Furthermore, compared with vehicle, treatment with T0901317 limited pyolysin-induced changes in cell shape (83 ± 5% vs 37 ± 9% cells damaged, t-test, n = 3, P < 0.05, Figure 6D). We then A B FIGURE 5 | Potassium and calcium do not affect 27-hydroxycholesterol cytoprotection. HeLa cells were cultured for 24 h in serum-free medium with vehicle or 10 ng/ml 27-hydroxycholesterol , and then challenged with control medium (■) or 100 HU pyolysin (■) for 2 h in (A) low or high-potassium medium, or (B) in calcium-containing or calcium-free medium. The leakage of LDH was measured in cell supernatants and viable cells assessed by MTT assay. Data are presented as mean + s.e.m. with dots representing the values of cells from 4 independent experiments; statistical significance was determined using ANOVA with Tukey's post hoc test, ***P < 0.001, **P < 0.01, *P < 0.05. confirmed that siRNA targeting NR1H3 and NR1H2 diminished the protective effects of T0901317 ( Figure 6E). These data provide evidence that LXR activation protected cells against pyolysin, but that this was not a mechanism for 27hydroxycholesterol cytoprotection.

Oxysterols Reduce Accessible Cholesterol
We next explored other cholesterol regulation mechanisms that might explain hydroxycholesterol cytoprotection. We first examined whether culturing HeLa cells with additional cholesterol might diminish oxysterol cytoprotection, but the addition of 10% serum did not prevent treatment with 25hydroxycholesterol protecting against a subsequent pyolysin challenge ( Figure 7A). However, maximal cytoprotection required treatment with 100 ng/ml 25-hydroxycholesterol in the presence of serum, compared with 10 ng/ml in serum-free medium. As oxysterols have a similar structure to cholesterol, we also considered whether oxysterols might bind to pyolysin to prevent cytolysis. However, mixing pyolysin with 27- hydroxycholesterol or 25-hydroxycholesterol before challenging untreated cells did not diminish pyolysin-induced LDH-leakage or cytolysis ( Figure 7B). An additional mechanism that might be relevant is that 27hydroxycholesterol stimulates inactivation of HMGCR, although the expectation would be that total cellular cholesterol abundance should also be reduced (34,51). Inhibiting HMGCR with 10 µM atorvastatin reduced pyolysin-induced leakage of LDH and cytolysis (Supplementary Figure 6A) and reduced total cellular cholesterol (Supplementary Figure 6B).
As side-chain hydroxycholesterols did not alter total cellular cholesterol, we investigated cell membrane accessible cholesterol. Treatment with 27-hydroxycholesterol reduced pyolysin binding to HeLa cells by 71%, as determined using an anti-pyolysin antibody ( Figure 7D). Side-chain hydroxycholesterols stimulate ACAT esterification of cholesterol, which reduces accessible cholesterol in cell membranes (10,28,52,53). Culturing HeLa cells with the ACAT inhibitor SZ58-035 for 16 h before and during 24 h treatment with 27-hydroxycholesterol or 25hydroxycholesterol diminished cytoprotection against pyolysininduced LDH leakage and cytolysis ( Figure 7E). If cytoprotection depends on side-chain hydroxycholesterols reducing cell membrane accessible cholesterol, we reasoned that increasing this pool of cholesterol should also diminish cytoprotection. Accessible cholesterol can be generated using sphingomyelinase to release sphingomyelin-sequestered cholesterol (5,10). Treating HeLa cells with sphingomyelinase both increased the sensitivity of cells to pyolysin damage, and diminished oxysterol cytoprotection against pyolysin ( Figure 7F; P < 0.001). Together these data support the concept that sidechain hydroxycholesterols reduce accessible cholesterol in the cell membrane.

DISCUSSION
We found that oxysterols stimulated the intrinsic protection of epithelial cells against damage caused by cholesterol-dependent cytolysins. Treatment with 27-hydroxycholesterol protected four types of epithelial cells against a subsequent challenge with pyolysin, and protected cells against streptolysin O and Staphylococcus aureus a-hemolysin. Treatment with 27hydroxycholesterol reduced cytolysin-induced leakage of potassium ions and LDH, limited cytoskeletal changes, and reduced cytolysis. However, oxysterol cytoprotection did not appear to depend on potassium or calcium flux, or activation of LXRs. Instead, cytoprotection was dependent on ACAT reducing accessible cholesterol in cell membranes. Collectively, these findings imply that oxysterols stimulate the intrinsic protection of epithelial cells against pore-forming toxins and may help protect tissues against pathogenic bacteria. This oxysterol cytoprotection may help complement the role of oxysterols in modulating immune cell responses to bacteria (10,27). We reason that protecting epithelial cells against pore-forming toxins is likely to be more evolutionary and energetically advantageous than repairing damage or mounting inflammatory responses. Preventing damage also helps tissues and organisms better tolerate the presence of pathogens (11). Finally, oxysterol cytoprotection might inspire new prophylactic applications that limit tissue damage caused by bacteria. An important physiological role of the epithelium is to protect underlying tissue cells against bacterial infections (1,11,54). We found that pyolysin damaged several epithelial cell types, which is consistent with widespread effects of T. pyogenes and pyolysin across species and tissues (19)(20)(21)24). However, our main finding was that treating HeLa cells with either 27-hydroxycholesterol or 25hydroxycholesterol reduced pyolysin-induced cytolysis from > 68% to < 2%. There was also evidence of less pyolysin poreformation with reduced pyolysin binding to cells, less potassium and LDH leakage, and prevention of pyolysin-induced MAPK phosphorylation. Side-chain hydroxycholesterol cytoprotection against pyolysin extended to A549, Hep-G2, and NCI-H441 epithelial cells. Although 25-hydroxycholesterol had no significant effect on L. monocytogenes escape from phagocytic vacuoles, which requires the cholesterol-dependent cytolysin listeriolysin O (10), we found 27-hydroxycholesterol also protected cells against streptolysin O. Our findings add to recent observations that interferonstimulated production of 25-hydroxycholesterol protects murine macrophages against perfringolysin O, streptolysin O and anthrolysin O damage (28); and that 27-hydroxycholesterol or 25hydroxycholesterol protects bovine endometrial epithelial and stromal cells against pyolysin (29). However, unlike the previous studies, we found that the ring-modified hydroxycholesterols 7ahydroxycholesterol and 7b-hydroxycholesterol also partially protected HeLa cells against pyolysin, although they were not as potent as 27-hydroxycholesterol. Taking the evidence together, we suggest that oxysterols may contribute to maintaining or enhancing epithelial barriers against bacterial cytolysins. Further work is needed to determine the bioavailability of oxysterols in epithelial tissues and whether oxysterols regulate resilience to infection of epithelial barriers in vivo. However, oxysterols such as 27hydroxycholesterol and 25-hydroxycholesterol are present in ng/ ml amounts in the bovine endometrium, and 25-hydroxycholesterol is secreted by endometrial epithelial cells in response to lipopolysaccharide or pyolysin (29). Furthermore, injection of 25hydroxycholesterol into the skin of mice reduced the tissue damaged caused by a subsequent injection of anthrolysin O (28).
Eukaryotic cells have evolved several mechanisms to respond to damage caused by pore-forming toxins (14,16,33). Cell membrane and cytoskeletal repair mechanisms are triggered by potassium efflux from cells and increased intracellular calcium (14)(15)(16). We therefore considered whether side-chain hydroxycholesterols may act by enhancing these ion fluxes to trigger damage repair responses, rather than protecting the cells against poreformation. However, we did not find evidence that side-chain hydroxycholesterol cytoprotection was associated with changes in potassium or calcium. Thus, we suggest that hydroxycholesterols enhance the intrinsic protection of cells against damage, rather than improving their ability to recover from damage.
Reducing or modifying cell membrane cholesterol is an obvious mechanism for altering the intrinsic protection of cells against cholesterol-dependent cytolysins (3,10,21,24,26). There are three pools of cell membrane cholesterol: an essential pool, a sphingomyelin-sequestered pool, and a labile pool of accessible cholesterol that can be bound by cholesterol-dependent cytolysins (5,25). Activation of LXRs can stimulate cholesterol efflux from cells (49), and the LXR agonist T0901317 partially protected HeLa cells against pyolysin in the present study. Similarly, another synthetic LXR agonist, GW3965, also protects macrophages against cholesterol-dependent cytolysin damage (28). Although, side-chain hydroxycholesterols are LXR agonists (49,55), using siRNA to target NR1H2 and NR1H3 did not diminish oxysterol cytoprotection in the present study. At micromolar concentrations, 27-hydroxycholesterol can also act via estrogen receptors, but this mechanism is unlikely to be important for the present study because estrogen receptors are not expressed by HeLa or A549 cells (56,57). We also considered whether 27-hydroxycholesterol inactivation of HMGCR might be important because inhibiting HMGCR protects bovine endometrial stromal cells against pyolysin (34). Surprisingly, 27-hydroxycholesterol and 25-hydroxycholesterol did not reduce total cellular cholesterol, as expected when inhibiting HMGCR (34,51). However, this observation concurs with a similar finding using Chinese hamster ovary cells treated with 2 µg/ml 25-hydroxycholesterol for 20 h (10). Instead, we found that 27-hydroxycholesterol reduced pyolysin binding to cells, which is consistent with less accessible cholesterol in the cell membrane (4). Side-chain hydroxycholesterols reduce accessible cholesterol in cell membranes by stimulating ACAT esterification of cholesterol (10,28,53). In agreement with this concept, oxysterol cytoprotection against pyolysin was diminished in the present study by inhibiting cholesterol esterification or by releasing sphingomyelin-bound cholesterol. However, as some cytoprotection against pyolysin remained in these experiments, oxysterol-stimulated ACAT cholesterol esterifi cation may not be the sole mechanism for cytoprotection. For example, we cannot rule out the possibility that oxysterols may protect against cytolysins by inhibiting cholesterol biosynthesis as a result of oxysterol binding to INSIG to prevent processing of SREB2 (26).
It was intriguing that treatment with 27-hydroxycholesterol also protected HeLa and A549 cells against Staphylococcus aureus a-hemolysin. A longer challenge period was required to induce a-hemolysin cytolysis compared with pyolysin, possibly due to the ten-fold smaller pore-diameter (22,46). Conversely, these smaller pores have a lower calcium permeability than cholesterol-dependent cytolysins, which can impair calciumdependent protective responses (58,59). A further benefit of oxysterols for limiting the severity of disease is that oxysterols can also suppresses macrophage inflammatory response to pathogens (8,27). Therefore, future in vivo studies might test the prophylactic effect of oxysterol treatment on the severity of Staphylococcus aureus infection, as well as infections caused by bacteria producing cholesterol-dependent cytolysins.
In conclusion, we found that oxysterols stimulated the intrinsic protection of epithelial cells against damage caused by cholesterol-dependent cytolysins. In particular, treatment with 27-hydroxycholesterol protected HeLa cells against a subsequent challenge with pyolysin or streptolysin O. Oxysterol cytoprotection extended to other cell types and protected against Staphylococcus aureus a-hemolysin. Mechanistically, oxysterol cytoprotection was at least partially dependent on ACAT reducing accessible cholesterol in the cell membrane.
Collectively, these findings imply that oxysterols stimulate the intrinsic protection of epithelial cells against pore-forming toxins and may help protect tissues against pathogenic bacteria.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.