Sphingosine kills intracellular Pseudomonas aeruginosa and Staphylococcus aureus

Abstract

Sphingosine has been previously shown to kill many strains of pathogenic bacteria including Pseudomonas aeruginosa, Staphyloccus aureus, Acinetobacter, and atypical mycobacteria. However, these studies were performed on isolated or extracellular bacteria and it is unknown whether sphingosine also targets intracellular bacteria. Here, we demonstrate that exogenously-added sphingosine directly binds to extracellular P. aeruginosa and S. aureus, but also targets and binds to intracellular bacteria. Intracellular sphingosine and bacteria were identified by sequential immunostainings. We further show that exogenously-added sphingosine also kills intracellular P. aeruginosa and S. aureus using modified gentamycin assays. Intracellular killing of P. aeruginosa and S. aureus by sphingosine is not mediated by improved phagosomal-lysosomal fusion. In summary, our data indicate that sphingosine binds to and most likely also directly kills extra- and intracellular P. aeruginosa and S. aureus.

Sphingosine directly targets and kills extra- and intracellular bacteria.

Introduction

Cystic fibrosis (CF) is caused by mutations of the cystic fibrosis transmembrane conductance regulator (human: CFTR, murine: Cftr) (Elborn 2016). CF is one of the most common autosomal recessive disorders in the EU and the USA. Statistically, 1 child out of 2500 births is affected by cystic fibrosis (Elborn 2016). This results in approximately 80 000 individuals suffering from CF in the EU and a similar number in the USA (Elborn 2016). Gastrointestinal problems caused by CFTR-deficiency determined life expectancy until the development and application of digestive enzymes. These problems are now usually well controlled. At present, pulmonary infections and pulmonary complications are clinically the most important problems for CF patients and also determine the life expectancy of the patients. CF patients often suffer from chronic pulmonary inflammation, lung fibrosis, and, most importantly, recurrent and chronic infections with Staphylococcus aureus (S. aureus) Pseudomonas aeruginosa (P. aeruginosa) Burkholderia cepacia, Haemophilis influenzae, atypical mycobacteria and other bacteria (CFF Patient Registry Annual Data Report).

We and others have shown that CF epithelial cells exhibit a marked imbalance of ceramide and sphingosine (Teichgräber et al. 2008, Becker et al. 2010, Brodlie et al. 2010, Bodas et al. 2011, Caretti et al. 2014, Pewzner-Jung et al. 2014, Tavakoli Tabazavareh et al. 2016, Grassmé et al. 2017, Becker et al. 2020, Gardner et al. 2020, Liessi et al. 2020, Loberto et al. 2020). Bronchial epithelial cells from CF mice, biopsies from human nasal polyps, human nasal epithelial cells, and transplant specimens from individuals with CF exhibit increased levels of specific ceramide species, while sphingosine levels are greatly decreased in CF epithelial cells (Teichgräber et al. 2008, Becker et al. 2010, Brodlie et al. 2010, Bodas et al. 2011, Caretti et al. 2014, Pewzner-Jung et al. 2014, Tavakoli Tabazavareh et al. 2016, Grassmé et al. 2017, Becker et al. 2020, Gardner et al. 2020, Liessi et al. 2020, Loberto et al. 2020). In contrast, sphingosine is abundantly expressed on the luminal surface of healthy humans and murine nasal, tracheal, and bronchial healthy epithelial cells, while ceramide levels are relatively low in these cells of healthy mice or humans (Pewzner-Jung et al. 2014, Tavakoli Tabazavareh et al. 2016, Grassmé et al. 2017, Becker et al. 2020). We have shown that the enzyme acid ceramidase, which converts ceramide into sphingosine, is down-regulated in CF cells and that this lack of acid ceramidase expression causes the imbalance between ceramide and sphingosine (Grassmé et al. 2017, Becker et al. 2020, Gardner et al. 2020). The down-regulation of sphingosine in CF airways seems to cause the high infection susceptibility of CF mice and CF patients since (i) sphingosine kills many bacterial species in vitro and in vivo very efficiently, including P. aeruginosa, Staphylococcus aureus (S. aureus), Staphylococcus epidermidis, Haemophilus influenzae, Escherichia coli, Moraxella catarrhalis, Burkholderia cepacia, and Acinetobacter baumanii (Bibel et al. 1992, Fischer et al. 2013, Pewzner-Jung et al. 2014, Tavakoli Tabazavareh et al. 2016, Grassmé et al. 2017, Azuma et al. 2018, Carstens et al. 2019, Seitz et al. 2019, Becker et al. 2020, Verhaegh et al. 2020) and (ii) inhalation of sphingosine or acid ceramidase by CF mice eliminated and prevented pulmonary P. aeruginosa and S. aureus infections without side effects (Pewzner-Jung et al. 2014, Tavakoli Tabazavareh et al. 2016, Grassmé et al. 2017, Becker et al. 2020).

The anti-bacterial role of sphingosine in airways was recently also shown in isolated perfused and ventilated pig lungs. These studies revealed that sphingosine inhalation is sufficient to eliminate intrabronchial pathogens without visible adverse effects on epithelial cells of the respiratory tract (Carstens et al. 2019, Carstens et al. 2021, Liu et al. 2024).

Mechanistically, we demonstrated that sphingosine binds to bacteria and alters the properties such as fluidity of membranes, very likely by crosslinking large lipids such as cardiolipin (Verhaegh et al. 2020). These alterations result in rapid leakiness of the bacteria and their death (Verhaegh et al. 2020), although other mechanisms were not excluded by these studies.

Collectively, these data indicate that sphingosine kills extracellular bacteria. At present, it is unknown whether sphingosine also affects intracellular bacteria.

Thus, we investigated the effects of exogenously added sphingosine on intracellular pathogens. Our data demonstrate that infection of epithelial cells with S. aureus or P. aeruginosa results in internalization of the bacteria. Addition of sphingosine or blockade of sphingosine consumption by treatment with sphingosine kinase inhibitors resulted in death of intracellular S. aureus and P. aeruginosa. The use of biotinylated sphingosine suggested that sphingosine directly binds to and kills intracellular bacteria.

Methods

Bacteria preparation

We used two distinct strains of Staphylococcus aureus, i.e. a clinical isolate named DH, and the laboratory strain RN 6930 and the Pseudomonas aeruginosa strains ATCC 27853, a laboratory strain, and a clinical P. aeruginosa isolate, named 696. The P. aeruginosa strain 696 is a very invasive, strain, but shows almost no acute toxicity. Bacteria were cultured from frozen stocks on tryptic soy broth agar plates overnight. Bacteria were then removed from the agar plate and transferred to tryptic soy broth (TSB). The optical density of bacterial suspension TSB was adjusted to 0.2–0.25 and the samples were grown at 37°C with shaking at 125 rpm for 1 hour for the P. aeruginosa strains and 75 min for the S. aureus strains to achieve reproducible growth conditions at the early logarithmic phase and thereby to ensure consistent infection conditions. Bacteria were then washed twice in HEPES/Saline (20 mM HEPES, 132 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 0.7 mM MgCl₂, 0.8 mM MgSO₄, pH 7.4) by centrifugation at 3200 rpm for S. aureus and 2800 rpm for all P. aeruginosa strains for 10 min and finally suspended in H/S for subsequent assays.

Sphingosine preparation

We used sphingosine (Avanti Polar Lipids, #860490P) dissolved in n-octyl-β-D-glucopyranoside (OGP) at a concentration of 20 mM or 10 mM. Before use, we sonicated the sphingosine solution for 10 min in a bath sonicator (Bandelin, Sonorex) and then diluted it in H/S or RPMI-1640 (Gibco) supplemented with 10 mM HEPES (pH 7.4) and 0.5% fetal calf serum (FCS) to reach the intended concentrations of 1, 5, 10, or 20 µM sphingosine. The diluted samples were again sonicated for 10 min immediately prior to use.

In vitro sphingosine killing effect

To determine the effects of sphingosine on S. aureus in vitro, we incubated 5 × 10⁵ colony-forming units (CFU) of S. aureus strains DH or 6390 with 5 µM, 10 µM, or 20 µM sphingosine in 1 mL RPMI 1640 supplemented with 10 mM 4-(2-hydroxyethyl)-1-piperazineethane-sulfonic acid (HEPES) and 0.5% fetal calf serum (FCS) for 2 h. Controls were left untreated or incubated with the corresponding concentration of ocytylglucopyranoside. After washing the bacteria, we spread an aliquot of the bacterial solution onto Luria broth (LB) agar plates. We incubated the plates overnight to allow the bacteria to grow and form visible colonies, which we then counted to determine the number of CFUs.

In addition, we modified the conditions and tested whether sphingosine also kills the bacteria within a shorter time. To this end, we incubated 10.000 CFU of P. aeruginosa strains ATCC 27853, or 696 or of the S. aureus strain DH in 500 µL H/S (pH 7.0) for 45 min in the presence or absence of 10 µM or 20 µM sphingosine, diluted in H/S, plated aliquots on LB agar plates and determined the remaining CFU after o/n growth.

Chang cells

To investigate the effects of sphingosine on intracellular S. aureus and P. aeruginosa, we used Chang cells, an immortalized conjunctiva epithelial cell line of human origin. Cells were grown in RPMI 1640 medium (Gibco) supplemented with 10% FCS, 10 mM HEPES, 2 mM L-glutamine, 1 mM sodium pyruvate, 100 µM nonessential amino acids, and 100 U/mL penicillin and 100 µg/mL streptomycin (all Gibco) at 37°C in a 5% CO₂ with humidified conditions.

Cellular Infections

We seeded 50 000 Chang cells in 24 well plates 24 h before infection. On the day of infection, we replaced the medium with RPMI 1640 supplemented with 10 mM HEPES (pH 7.4). We then infected the cells at a multiplicity of infection (MOI) of 1 cell : 10 bacteria for 4 h to allow for the bacteria to interact with the host cells and to become internalized. Then, we extensively washed the samples to remove most of the extracellular bacteria and added sphingosine to the medium at varying concentrations of 1, 5, or 10 µM. Finally, we carefully removed the medium and collected the infected Chang cells for further analysis.

Alternatively, we seeded 30 000 Chang cells in a 24 well plate 24 h prior to infection, washed in H/S (pH 7.0) and infected the cells with P. aeruginosa ATCC 27853 or 696 in H/S (pH 7.0) for 30 min. Infection was initiated by addition of the bacteria at an MOI of 1 cell : 100 bacteria and a short centrifugation of the plates at 1000 rpm for 2 min. After the infection, the medium was removed, and the samples were treated for 45 min with 10 µM or 20 µM sphingosine in H/S (pH 7.0). In order to test the effects of cytochalasin D (Sigma) on killing of intracellular bacteria with sphingosine, Chang cells were infected with P. aeruginosa ATCC 27853 or 696 for 30 min followed by a 60 min incubation with 10 µM cytochalasin D to block the cytoskeleton, followed by an additional 45 min incubation with 10 µM sphingosine in PBS (pH 7.0). Cytochalasin was left on the cells during the incubation with sphingosine. Controls were treated for the same times but no cytochalasin D was added. Samples were then processed as described below to determine the remaining intracellular bacteria.

Intracellular bacteria killing assay

To determine the number of intracellular bacteria, Chang cells were infected with the P. aeruginosa strains described above. After infection the cells were extensively washed with PBS to remove most of the extracellular bacteria, treated with sphingosine, and then incubated with polymyxin (100 µg/mL) to kill extracellular P. aeruginosa. Polymixin is not cell permeable and thereby only kills extracellular bacteria, but not the intracellular bacteria. Samples were incubated for 30 min with the antibiotics, washed, and then treated with 0.5% saponin to permeabilize the cell membrane, and the resulting lysate was collected and plated on agar plates to determine the number of bacteria present. Saponin lyses the mammalian cell membrane but not the bacteria. The number of colony-forming units (CFU) was counted after o/n incubation, and the data was analyzed to determine the efficiency of the infection.

Alternatively, extracellular bacteria, in particular S. aureus, which are less adhesive to plastic surfaces than P. aeruginosa, were removed by washing the cells 4-times with PBS. Elimination of extracellular bacteria was then confirmed by counting the CFU in the last wash buffer.

In order to exclude that the observed effects of sphingosine on intracellular bacteria are overlaid by preventing internalization of bacteria, Chang cells were infected with the P. aeruginosa strains for 30 min, extensively washed with PBS, incubated with polymyxin (100 µg/mL) for 30 min to kill extracellular P. aeruginosa, washed extensively and then incubated with sphingosine.

Treatment of Chang cells with sphingosine kinase inhibitor I-II or the acid ceramidase inhibitor B13

50 000 Chang cells were seeded in 24 wells and treated with 5 µM sphingosine kinase inhibitor I-II (abcam, ab141594) or with the acid ceramidase inhibitor B13 (20 µM) (MEDCHEM Express, HY-124962) for 1 hour. Cells were then infected with S. aureus and treated with sphingosine, as described above. The number of intracellular bacteria was analyzed as above.

Immunofluorescence studies with biotin-sphingosine

Chang cells were seeded on coverslips in 24 wells. Cells were washed twice with H/S immediately prior to infection. Samples were then infected with P. aeruginosa strains ATCC 27853 or 696 for 30 min in H/S. To determine binding of sphingosine to extracellular bacteria, the samples were incubated for 5, 10, 15 or 30 min with 2 µM biotin-sphingosine (Cayman, #39 291) in H/S. Samples were washed in H/S, fixed in 1% PFA buffered at pH 7.3 for 10 min, washed, permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature, washed in PBS, blocked with 5% FCS in PBS for at least 10 min and incubated with Cy3-coupled streptavidin (1:500, Jackson Immunoresearch # 016–160-084) for 45 min at room temperature. Samples were washed 3-times with PBS plus 0.05% Tween 20 (Sigma) and once with PBS and embedded in Moviol. Samples were analyzed using a Leica LEICA TCS SL confocal microscope with a 40x or 100x lens. These studies served to determine the kinetics of the uptake of biotin into bacteria, but do not really allow to discriminate intracellular from extracellular bacteria.

Immunofluorescence studies to detect binding of sphingosine to intracellular bacteria

To determine sphingosine in intracellular P. aeruginosa, we infected Chang cells with the P. aeruginosa strains ATCC27853 or 696 for 30 min in H/S and incubated with 2 µM FITC-sphingosine (Echelon Biosciences, #E00253-025-23), diluted in H/S (pH 7.0) for 45 min. Samples were washed in PBS, fixed in 1% PFA buffered at pH 7.3 for 10 min, washed and stained as following: Extracellular bacteria were stained with rabbit anti-P. aeruginosa antibodies (1:2000, Abcam # ab68538) for 30 min in PBS at room temperature. Samples were washed 3-times in PBS and stained with AlexaFluor 647-coupled F(ab)₂ donkey anti-rabbit fragments (1:1000, Jackson Immunoresearch, #711–606-152) for 30 min. Cells were washed again 3-times in PBS, permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature, washed in PBS, blocked with 5% FCS in PBS for at least 10 min and stained again with anti-P. aeruginosa antibodies as above (1:2000) for 30 min in PBS at room temperature, washed and then incubated with Cy3-coupled F(ab)₂ donkey anti-rabbit fragments (1:1000, Jackson Immunoresearch, #711–166-152) for 30 min. Samples were finally embedded in Moviol. Samples were analyzed using a Leica LEICA TCS SL confocal microscope with a 40x or 100x lens.

Extracellular bacteria should be stained blue (Alexa Fluor 647), red (Cy3) and green (FITC) if they bound sphingosine; intracellular bacteria should be stained green (FITC) for sphingosine and red (Cy3) for P. aeruginosa.

Statistical analysis and quantification

Data are given as mean ± standard deviation. The statistical analysis was conducted using GraphPad Prism software, with Student's t-test for single comparisons and ANOVA for multiple comparisons followed by post hoc Bonferroni's multiple comparison test. Results with a P-value of less than 0.05 were considered statistically significant.

Results

To determine whether exogenously added sphingosine affects intracellular bacteria, we first confirmed the effect of exogenously added sphingosine on S. aureus strain DH, a clinical strain isolated from a patient with a S. aureus sepsis, on ATCC 27853, a laboratory P. aeruginosa strain, and on P. aeruginosa strain 696, a very invasive, but less toxic strain. Since we have previously shown that the effects of sphingosine depend on protonation of the NH2 group, we performed the experiments in H/S at pH 7.0, which also mimics the pH on the surface of the airways (Verhaegh et al. 2020). These studies confirm previous data and show rapid killing of P. aeruginosa strains ATCC 27853 and 696 as well as of S. aureus DH in H/S by sphingosine (Fig. 1A).

Figure 1.

Sphingosine kills Pseudomonas aeruginosa and Staphylococcus aureus strains in vitro under different conditions P. aeruginosa strains ATCC 27853 or 696 (A) or S. aureus strains DH or 6930 (B) were incubated for 30 min with 20 µM sphingosine in H/S or left untreated (A) or for 2 h with 5, 10, or 20 µM sphingosine in RPMI 1640 supplemented with 10 mM HEPES and 0.5% fetal calf serum or left untreated (B, C). Bacterial numbers were determined by counting colonies after growth o/n on LB agar plates. Given are the mean ± SD of 6 (A), 14 (B) or 5 (C) independent experiments; ***P < 0.001 ANOVA and post hoc Tukey´s multiple comparison test.

We also modified the infection conditions and incubated the bacteria in RPMI-1640 buffered to pH 7.4 in the presence of 0.5% FCS, which will bind and presumably neutralize some of the sphingosine. We therefore extended the incubation time to allow sufficient sphingosine to reach the bacteria. These studies also revealed efficient killing of S. aureus DH and P. aeruginosa ATCC 27853 with approximately 90% killing of the bacteria after 2 h incubation with 10 µM sphingosine and almost complete killing of the bacteria with 20 µM sphingosine (Fig. 1B). A longer exposure of the bacteria to sphingosine was necessary in these experiments to achieve killing of the bacteria as in the initial experiments performed in H/S, very likely because of the presence of albumin and other proteins in fetal calf serum.

To confirm these data on a different strain, we used the S. aureus strain 6930 and treated isolated bacteria with 5, 10, or 20 µM sphingosine for 2 h. The results indicate that even low concentrations of sphingosine killed S. aureus strain 6930, with the death of approximately 90% of the bacteria at 5, 10, or 20 µM sphingosine (Fig. 1C).

Next, we tested whether sphingosine also kills intracellular bacteria. To this end, we infected Chang epithelial cells for 30 min with P. aeruginosa strains ATCC 27853 or 696 in H/S, pH 7.0. We then added sphingosine at 10 µM concentration in H/S, pH 7.0, and incubated the samples for 45 min. Samples were washed, extracellular bacteria were killed by polymyxin, the cells were lysed in saponin, the lysates were plated, and the number of intracellular bacteria was determined after o/n growth. These studies revealed that exogenous sphingosine also killed intracellular P. aeruginosa (Fig. 2A).

Figure 2.

Sphingosine kills and directly binds to intracellular P. aeruginosa (A) Chang cells were infected with P. aeruginosa strain ATCC 27853 for 30 min and treated with 10 µM sphingosine for 45 min. Cells were washed, extracellular bacteria were killed by incubation with polymixin, washed, lysed and the number of intracellular bacteria was determined by counting colony forming units after plating aliquots on LB plates and o/n growth. Sphingosine added after incubation with polymyxin still reduced the number of intracellular bacteria. (B) Incubation of infected cells with cytochalasin prior to treatment with sphingosine does not alter the effect of sphingosine on the bacteria. (C,D) Incubation of infected cells with biotin-sphingosine followed by staining with Cy3-coupled streptavidin shows direct binding of sphingosine to P. aeruginosa. Panel C displays the confocal microscopy analysis upon subtraction of the background due to the presence of endogenous biotin in the bacteria. Panel D shows an optimized exposure time. Arrows indicate intracellular bacteria as defined by the presence of a vacuole around the bacteria and the presence of the bacteria in the same focus level as cellular structures. (E-G) Chang cells were infected with P. aeruginosa (P.a.) strains ATCC 27853 or 696 for 30 min to allow internalization of the bacteria. Infected cells were then treated with FITC-sphingosine for 45 min, washed, fixed, stained with Alexa Fluor 647-coupled anti-P. aeruginosa antibodies, washed, permeabilized, blocked and stained with Cy3-coupled anti-P. aeruginosa antibodies. Extracellular bacteria that bind FITC-sphingosine will then appear in purple (green + blue + red), intracellular bacteria in yellow (green + red). Please note that FITC sphingosine certainly also binds to cellular membranes. Panel E and F stainings for P. aeruginosa ATCC 27853 (two examples) and panel G for P. aeruginosa 696, which is very invasive and no extracellular bacteria were present. Arrows indicate intracellular bacteria. (H) Control experiments confirm that biotin-sphingosine and FITC-sphingosine kill P. aeruginosa, very similar to unmodified sphingosine. Shown are the mean ± SD of 6 independent experiments (A, B, H); ***P < 0.001 ANOVA and post hoc Tukey´s multiple comparison test. Panel C-G show representative stainings from 6 independent experiments.

In order to exclude that the observed effects of sphingosine on intracellular bacteria are overlaid by preventing internalization of bacteria, Chang cells were infected with the P. aeruginosa strains for 30 min, extensively washed with PBS, incubated with polymyxin and then incubated with sphingosine. The results show that sphingosine added after incubation with polymyxin still reduced the number of intracellular bacteria (Fig. 2A).

To test whether sphingosine indirectly induces killing of intracellular bacteria, for instance, by promoting fusion of the phagosomes with lysosomes, or whether sphingosine directly kills the bacteria, we infected with P. aeruginosa ATCC 27853 and let the bacteria invade the cells. We then incubated Chang cells for 60 min with cytochalasin D, which blocks the cytoskeleton, and finally treated with 10 µM sphingosine. These studies revealed that pre-incubation of Chang cells with Cytochalasin D did not change killing of intracellular P. aeruginosa by sphingosine (Fig. 2B), suggesting that sphingosine does not kill the pathogens by promoting phagolysosomal fusion.

To test whether sphingosine binds to the bacteria, we infected Chang cells for 30 min with P. aeruginosa ATCC 27853 and then incubated with biotin-sphingosine for 5, 10, 15 or 30 min, fixed the samples in paraformaldehyde, permeabilized the cells and stained with Cy3-labelled streptavidin. The samples were washed, embedded in Mowiol, and analyzed by confocal microscopy. The results reveal a rapid binding of exogenously added biotin-sphingosine to internalized and extracellular bacteria (Fig. 2C), although these studies do not allow a perfect discrimination between intracellular and extracellular bacteria (please see below). Here, intracellular bacteria were identified by their presence in a vacuole and by the focus level. Since bacteria endogenously contain some biotin, the stainings showed some (expected) background staining. To eliminate this background, we set the intensity of the laser in the confocal microscope to a level that resulted in abrogation of the background staining. We then analyzed the biotin-sphingosine-incubated bacteria at exactly the same conditions (Fig. 2C). These studies revealed a binding of sphingosine to the bacteria, but they do not allow to detect the Chang cells anymore. Chang cells certainly also contain some endogenous biotin and will exhibit some staining, which can be employed to visualize the cells. We, therefore, collected a 2^nd set of pictures with higher laser intensity (Fig. 2D), which shows (i) that intra- and extracellular bacteria bind biotin-sphingosine and (ii) that the labeling of intra- and extracellular bacteria with sphingosine is of similar intensities (Fig. 2D). These data indicate that sphingosine directly binds to intracellular P. aeruginosa.

To prove that sphingosine binds to intracellular bacteria, we performed a sequential staining. To this end, Chang cells were infected with P. aeruginosa strains ATCC 27853 or 696, treated with FITC-sphingosine, fixed and washed. The samples were then stained with Alexa Fluor 647-coupled anti-P. aeruginosa antibodies prior to permeabilization of the cells. After staining the extracellular bacteria, the cells were permeabilized and stained with Cy3-coupled anti-P. aeruginosa antibodies.

These studies (Fig. 2E–G) revealed a binding of exogenously-added sphingosine to intracellular bacteria. Controls revealed that biotin-sphingosine and FITC-sphingosine still kill P. aeruginosa.

Next, we infected human conjunctiva Chang epithelial cells with S. aureus strain DH or S. aureus 6930 for 4 h, washed and then treated the cells for 2 h with 1, 5 and 10 µM sphingosine. The results demonstrate that 1, 5, and 10 µM sphingosine killed intracellular S. aureus, with approximately 50% killing of intracellular S. aureus strains DH and 6930 at 10 µM sphingosine (Fig. 3A,B).

Figure 3.

Sphingosine kills intracellular Staphylococcus aureus (A,B) Chang cells were infected with S. aureus DH or (C) S. aureus strain 6930 for 2 hours, then treated with sphingosine at the indicated concentrations for another 2 hours or left untreated. Bacterial numbers were determined by colony forming assay (A,C) or immunofluorescence staining (B). Data are presented as mean ± SD; n = 5–20. Statistical significance was determined by one-way ANOVA followed by post hoc Bonferroni's multiple comparison test tests; **p < 0.01, ^****p < 0001.

To test the effects of further lipids on the killing of bacteria, we treated S. aureus strain DH with ceramide, the precursor of sphingosine, sphingosine 1-phosphate and hexadecenal, the conversion products of sphingosine. C16 ceramide and sphingosine 1-phosphate did not affect the S. aureus strain DH (Fig. 4A,B). Incubation of S. aureus with 50 µM or 100 µM hexadecenal killed up to approximately 60% of the bacteria, but lower concentrations were without significant effect (Fig. 4C). These data indicate that neither ceramide nor sphingosine 1-phosphate or hexadecenal at physiological concentrations have an impact on S. aureus.

Figure 4.

Antimicrobial activity of sphingosine derivatives against S. aureus in vitro The bacterial count of S. aureus DH was assessed using the colony-forming unit (CFU) assay after treating with ceramide (A), sphingosine-1-phosphate (B), or hexadecenal (C) for 2 hours at the indicated concentrations. Data are presented as mean ± SD; n = 5. Statistical significance was determined by one-way ANOVA followed by post hoc Bonferroni's multiple comparison test, with **P < 0.01, ^****P < 0001.

We next tested whether an increase in endogenous sphingosine might also promote killing of intracellular bacteria. To this end, we treated Chang cells with the sphingosine kinase inhibitor I-II and tested whether this promotes killing of S. aureus strain DH. The results show that blocking sphingosine kinases with a relatively low dose of the inhibitor was as efficient to kill intracellular S. aureus strain DH as application of exogenous sphingosine with killing of approximately 60% of the cells (Fig. 5). Addition of exogenous sphingosine together with the kinase inhibitor did not significantly increase killing of intracellular bacteria (Fig. 5). In addition, we tested the effect of an inhibition of the acid ceramidase by the inhibitor B13 on survival of intracellular bacteria. The data show that this inhibitor promotes intracellular survival of P. aeruginosa (Fig. 5). This is now shown in Fig. 5. Control stainings of the cells with Cy3-labelled anti-sphingosine antibodies confirmed the accumulation of sphingosine in the cells upon application of the inhibitor (not shown).

Figure 5.

Endogenous sphingosine promotes killing of intracellular S. aureus Sphingosine kinases were blocked by emplyoing the sphingosine kinase inhibitor I-II prior to infection with S. aureus strain DH. Acid ceramidase was inhibited by incubation of cells with the acid ceramidase inhibitor B13. Survival of intracellular bacteria was determined after washing away extracellular bacteria, cell lysis and overnight growth of aliquots in LB plates. Data are presented as mean ± SD; n = 5–8. Statistical significance was determined by one-way ANOVA followed by post hoc Bonferroni's multiple comparison test test, ^****P < 0.0001.

Discussion

Sphingosine has been previously shown to kill a variety of pathogens, including E. coli, P. aeruginosa, S. aureus, S. epidermidis, Acinetobacter baumannii, Haemophilus influenzae, Burkholderia species or Mycobacteria abscessus. However, these studies investigated the effect of sphingosine on extracellular bacteria, while it remained undefined whether exogenous sphingosine is also able to bind to and kill intracellular bacteria. Here, we tested this question and investigated whether sphingosine kills intracellular P. aeruginosa and S. aureus. In the 1^st of experiments, we infected epithelial cells with P. aeruginosa, killed the exctracellular bacteria employing a cell-impermeable antibiotic, i.e. polymyxin, washed and then treated the cells with sphingosine. The cells were then lysed to release internalized bacteria. These studies revealed that sphingosine kills intracellular P. aeruginosa. Subsequent experiments confirmed this notion also for S. aureus.

While these studies demonstrate that exogenous sphingosine added to already infected cells kills intracellular bacteria, they do not answer the question whether sphingosine directly binds intracellular bacteria or whether sphingosine induces an indirect killing of intracellular bacteria by triggering for instance fusion of phagosomes with lysosomes. Our studies revealed a direct binding of biotin-sphingosine or FITC-sphingosine to the bacteria. Further, inhibition of the cytoskeleton using cytochalasin B did not alter the bactericidal effect of sphingosine. Collectively these data suggest that sphingosine acts directly on internalized bacteria. However, detailed immune electron microscopy studies investigating a direct binding of sphingosine to the internalized bacteria and the concomitant alteration of the morphology of intracellular bacteria seem to be required to unambiguously demonstrate a direct killing of intracellular bacteria by sphingosine. These very time consuming studies require the development of novel reagents, in particular antibodies detecting sphingosine, and therefore seem to be beyond the focus of the present study.

To address the kinetics of the binding of sphingosine to intra- and extracellular bacteria we performed detailed kinetic studies, which revealed that the binding of sphingosine to extracellular bacteria occured rapidly, i.e. within 1–5 minutes, reaches a maximum, and thereafter declines, while the binding of sphingosine to intracellular bacteria is a slower process with a maximum 20–30 min after addition of sphingosine. The reason for the decline of sphingosine binding to extracellular bacteria is presently unknown and needs to be defined in future studies.

We have previously shown that sphingosine binds to cardiolipin and induces a massive change of the viscosity of membranes (Verhaegh et al. 2020). The increased rigidity of the membrane upon binding of sphingosine resulted in a very rapid and massive increase of bacterial membrane permeability, loss of ATP and bacterial cell death (Verhaegh et al. 2020). Since mammalian plasma membranes do not contain cardiolipin, the effect was rather specific for bacteria. However, since sphingosine also binds to intracellular bacteria, it may also reach mitochondria, which contain cardiolipin and binding of sphingosine to mitochondrial cardiolipin may result in cell death. However, we did not observe death of Chang epithelial cells upon treatment of sphingosine, which might be explained by the expression of sphingosine kinase 2 in mitochondria (Strub et al. 2011) that may rapidly phosphorylate sphingosine to sphingosine-1-phosphate within the mammalian cell, which is not toxic, neither to mitochondria nor to bacteria. This is consistent with previous findings that demonstrated a very short half life time, i.e. within minutes, of sphingosine in mitochondria in Hela cells (Feng et al. 2018). Bacteria, at least most bacteria, do not express sphingosine kinases and this may explain their sensitivity to sphingosine.

Infection of epithelial cells with P. aeruginosa may also trigger some endogenous sphingosine, which may contribute to the killing of P. aeruginosa. However, in the experiments that employed addition of directly labeled biotin- or FITC-sphingosine to visualize sphingosine binding to internalized bacteria any endogenously-formed sphingosine does not interfere with the visualization of sphingosine, although it may reduce free binding sites for exogenously-added biotin- or FITC-sphingosine.

Our data suggest that sphingosine accumulates in intracellular bacteria. It is unknown whether sphingosine binds to specific targets, for instance cardiolipin, in bacteria and is therefore enriched in bacteria or whether it simply accumulates because bacteria do not express enyzmes to consume it into other products.

In summary, our data show that exogenously added sphingosine binds to and kills intracellular P. aeruginosa and S. aureus. Since both bacteria seem to be present within bronchial epithelial cells in lungs of patients with cystic fibrosis, inhalation of sphingosine may not only kill extracellular bacteria, but also the intracellular reservoir of the pathogens and may serve as a novel treatment of bacterial infections in cystic fibrosis lungs.

Author contributions

Helene May (Conceptualization, Data curation, Formal analysis, Investigation, Validation, Writing – original draft, Writing – review & editing), Yongjie Liu (Data curation, Formal analysis, Validation, Writing – review & editing), Stephanie Kadow (Data curation, Methodology, Supervision), Michael J. Edwards (Conceptualization, Formal analysis, Writing – review & editing), Simone Keitsch (Data curation, Investigation, Writing – review & editing), Barbara Wilker (Data curation, Investigation, Writing – review & editing), Markus Kamler (Conceptualization, Project administration, Supervision, Writing – review & editing), Heike Grassmé (Conceptualization, Investigation, Writing – review & editing), Yuqing Wu (Conceptualization, Data curation, Investigation, Writing – review & editing), and Erich Gulbins (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing)

Conflicts of interest

None declared.