Sphingoid long chain bases prevent lung infection by Pseudomonas aeruginosa

Cystic fibrosis patients and patients with chronic obstructive pulmonary disease, trauma, burn wound, or patients requiring ventilation are susceptible to severe pulmonary infection by Pseudomonas aeruginosa. Physiological innate defense mechanisms against this pathogen, and their alterations in lung diseases, are for the most part unknown. We now demonstrate a role for the sphingoid long chain base, sphingosine, in determining susceptibility to lung infection by P. aeruginosa. Tracheal and bronchial sphingosine levels were significantly reduced in tissues from cystic fibrosis patients and from cystic fibrosis mouse models due to reduced activity of acid ceramidase, which generates sphingosine from ceramide. Inhalation of mice with sphingosine, with a sphingosine analog, FTY720, or with acid ceramidase rescued susceptible mice from infection. Our data suggest that luminal sphingosine in tracheal and bronchial epithelial cells prevents pulmonary P. aeruginosa infection in normal individuals, paving the way for novel therapeutic paradigms based on inhalation of acid ceramidase or of sphingoid long chain bases in lung infection.


Introduction
Cystic fibrosis (CF) is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene and is found with a prevalence of 1 in 2,500-3,500 Caucasian newborns. CF patients, and patients with chronic obstructive pulmonary disease (COPD), are particularly susceptible to Pseudomonas aeruginosa infection, with approximately 80% of CF patients suffering from chronic P. aeruginosa pneumonia by 25 years of age, and one-third of patients with COPD hosting the bacteria (Rakhimova et al, 2009;Cystic Fibrosis Foundation, 2011). However, the increasing rate of P. aeruginosa resistance to many antibiotics necessitates the development of alternative therapies, which might also be beneficial in preventing P. aeruginosa infection in patients with trauma, burn wounds, sepsis, or in patients requiring ventilation (McManus et al, 1985;Crouch Brewer et al, 1996;Vidal et al, 1996).
Previous studies have demonstrated that various natural lipids and lyso-lipids act as bactericidal agents in skin (Fischer et al, 2012). Among these is the sphingoid long chain base (LCB), sphingosine (SPH), which protects human skin from bacterial colonization (Bibel et al, 1992). SPH is generated by hydrolysis of ceramide via acid ceramidase (AC). We now demonstrate that tracheal and bronchial epithelial SPH levels play a vital role in preventing normal individuals from P. aeruginosa lung infection. SPH levels are significantly reduced in tracheal and bronchial epithelia of CF patients and of CF mice, due to reduced AC activity, and normalization of SPH levels reverses susceptibility to P. aeruginosa. These data suggest two novel paradigms, namely that epithelial SPH is a natural antibactericidal agent in airways and that patients susceptible to bacterial infection in the lung could be treated by administration of AC or of SPH analogs.

Results and Discussion
Immunohistochemical analysis using an anti-SPH antibody demonstrated that SPH was abundantly expressed on the luminal surface of human nasal epithelial cells obtained from healthy individuals, but was almost undetectable on the surface of nasal epithelial cells from CF patients (Fig 1A -for validation of the specificity of the antibody, see Supplementary Information, Supplementary Fig S1  and Fig 2B). The reduction in SPH levels was recapitulated in tracheal and bronchial cells from CF mice (Fig 1B and C) and in airway epithelial cells from ceramide synthase 2 (CerS2)-null mice ( Fig 1C). CerS2-null mice also displayed increased ceramide levels specifically in tracheal and bronchial epithelial cells (Fig 1D and  Supplementary Fig S2), reminiscent of the elevated ceramide levels in these cells in CF mice and in CF patients ( Fig 1D) (Teichgräber et al, 2008;Becker et al, 2010;Brodlie et al, 2010;Ulrich et al, 2010;Bodas et al, 2011). Thus, CerS2-null mice display similar changes in ceramide and SPH levels as in CF mice despite their different genetic alterations. AC inhalation increased surface SPH in bronchial epithelial cells of CF and CerS2-null mice ( Fig 1C) with a concomitant reduction in ceramide levels ( Fig 1D).
To quantify SPH levels in the lung, we established a number of innovative methods, including: (i) mass spectrometry (MS) and (ii) enzymatic assays for SPH and ceramide using extracts from freshly isolated tracheal epithelial cells, which detects total SPH levels in these cells, (iii) in situ enzymatic assays for SPH and ceramide using the respective kinases applied directly on intact tracheal surfaces, which detects SPH and ceramide exclusively on the luminal surface, and (iv) immunoprecipitation of SPH upon incubation of the anti-SPH antibody at the luminal surface of intact trachea, which also detects SPH exclusively on the luminal surface. First, freshly isolated tracheal epithelial cells were extracted and SPH assayed using MS and enzymatic assays for SPH, demonstrating an approximately 75% reduction in total SPH levels in CF mice (Fig 2A). Next, an in situ enzyme assay, performed by application of SPH kinase (SK) and [ 32 P]cATP directly to the luminal side of the intact tracheal epithelial cell layer, revealed an approximately 75% reduction in SPH levels ( Fig 2B). The reduced SPH on the tracheal surface was confirmed by SPH immunoprecipitation using the anti-SPH antibody coupled to protein L-agarose beads, followed by lipid extraction and an enzymatic assay for SPH ( Fig 2B). Application of AC to the surface of isolated CF trachea prior to the in situ enzyme assay normalized SPH levels ( Fig 2B). Incubation of the isolated tracheal surface with 10 lM cytochalasin B (an actin filament polymerization inhibitor) prevented P. aeruginosa internalization into tracheal epithelial cells, but did not alter the amount of SPH detected by the in situ enzyme assay for SK or by SPH immunoprecipitation, excluding the possibility that SK and/or antibody internalization occurs during the assay (Fig 2B). These results demonstrate that SPH is present on the surface of WT epithelial cells while almost completely absent on the surface of CF epithelia.
We next demonstrated that AC or SPH inhalation increased SPH levels in CF tracheal epithelial cells and on the surface of CF trachea in vivo (Fig 2A and B).
Moreover, significant accumulation of ceramide was detected by mass spectrometry (MS) (Fig 2C, left) in extracts of isolated CF epithelial cells and by in situ kinase assay on the luminal surface of these cells in trachea of CF mice (Fig 2C, right), which was corrected by inhalation of AC ( Fig 2C). The specificity of the enzyme assay was confirmed by treating isolated trachea with AC in vitro (Fig 2C, right).
To determine the mechanism by which SPH levels are decreased on the surface of CF tracheal epithelial cells, AC activity was analyzed by loading trachea with [ 14 C]C16-ceramide and its consumption in vivo was analyzed. Significantly lower levels of AC activity were detected in CF mice ( Fig 2D). Cftr regulates the pH in secretory lysosomes (Barasch et al, 1991;Teichgräber et al, 2008) and probably also regulates the pH in small domains on the surface of tracheal epithelial cells that express proton pumps (Xu et al, 2012) by providing Clcounterions for H + , thus permitting continuous activity of protons pumps. Moreover, since alkalization results in a marked inhibition of AC activity (Li et al, 1998;Teichgräber et al, 2008), we next demonstrated that surface acidification of the trachea restored AC activity, increased surface sphingosine and decreased ceramide levels ( Fig 2D). To test whether acid-mediated hydrolysis of ceramide contributes to its consumption and the generation of sphingosine, we performed the experiment in the presence of the AC inhibitors, carmofur, or oleoylethanolamine. The inhibitors completely prevented sphingosine generation excluding significant acid-mediated hydrolysis of ceramide ( Fig 2D). It is important to note that the pH of small microdomains on the cell membrane of CF cells is independent of the airway liquid surface pH, which is decreased from pH 7.2 to pH 6.8 in CF cells (Pezzulo et al, 2012), which is probably not enough to alter AC activity.
Next, we examined whether sphingosine plays a role in infections with P. aeruginosa. We used three different P. aeruginosa strains to exclude any strain-specific effects. Upon intranasal infection, CF and CerS2-null mice displayed a dramatically increased sensitivity to P. aeruginosa strains 762 (Fig 3A), PA14, and ATCC 27853 ( Fig 3B), with 10-to 100-fold more bacteria in the lung 3-4 h after infection (Fig 3A and B), clinical signs of severe pneumonia ( Supplementary Fig S3A and B), massive release of cytokines such as interleukin-1b (IL-1b) ( Fig 3C) and TNF-a (Supplementary Fig S4) and influx of leukocytes ( Supplementary Fig S5A-H). Inhalation of CF mice with SPH to a level similar to that found in WT mice (Fig 2B), of AC or of FTY720 (a SPH analog in clinical use for treating multiple sclerosis) 1 h prior to P. aeruginosa infection, protected CF and CerS2-null mice from infection (  A-D Immunostaining of paraffin sections of nasal tissues from (A) healthy and CF individuals or (B) trachea and (C) bronchi of WT, CF, and CerS2-null mice using a Cy3coupled anti-SPH antibody. Effect of AC inhalation on SPH (C) and ceramide levels (D) in CF and CerS2-null mice. Representative images are shown; fluorescence levels were quantified and are given in arbitrary units (a.u.) (means AE s.d., n = 5 in A, n = 4 in B, n = 6 for WT controls for CF mice, n = 8 for CF, and n = 4 for all others in C, and n = 9 for untreated WT or CF, n = 7 for AC-inhaled WT, n = 8 for AC-inhaled CF and n = 4 for CerS2-null mice in D). Numbers above bars indicate the exact calculated P-values.  Ceramide (pmol/10 cells) . The low bacterial numbers and the lack of cytokines and leukocyte influx demonstrate that sphingosine and FTY720 directly and rapidly kill the bacteria, thus preventing further inflammation. Controls on non-infected mice showed that inhalation of SPH, AC, or FTY720 had no influence on pulmonary cytokines nor did it induce neutrophil influx 4 h, 1, 7, or 14 days after inhalation ( Fig 3C, Supplementary Figs S4, S5A, B, E, F, and S6A-C).

Control
P. aeruginosa was next incubated in vitro with a number of natural and synthetic sphingoid LCBs. A number of LCBs inhibited bacterial growth with EC 50 values of 0.3-2.2 lM, including the plant LCB, phytosphingosine, non-natural stereoisomers of SPH, and dihydrosphingosine (sphinganine), as well as LCBs which have recently been discovered to occur at low levels in mammals (Pruett et al, 2008) (Table 1). A number of other LCBs also inhibited bacterial growth with a somewhat higher EC 50 value, while other LCBs were without effect at concentrations as high as 50 lM (Table 1). These results demonstrate a notable specificity of the effect of LCBs on bacterial growth, providing structural information that could be used in development of novel LCB-based drugs to treat lung infection. To investigate whether other bacteria that affect CF patients are also susceptible to LCB treatment, we examined the effect of sphingosine on Acinetobacter baumannii, Haemophilus influenzae, Burkholderia cepacia, and Moraxella catarrhalis. Sphingosine inhibited growth of and/or killed Acinetobacter baumannii with an EC 50 of 0.07 AE 0.05 lM, of Moraxella catarrhalis with an EC 50 of 0.04 AE 0.004 lM of Haemophilus influenzae with an EC 50 of 4.8 AE 0.49 lM and Burkholderia cepacia with an EC 50 of 45 AE 6.3 lM. The relative resistance of Burkholderia cepacia might lie in its outer membrane unique composition (Cox & Wilkinson, 1991).
The results presented herein uncover a major innate defense mechanism of healthy airways, namely the bactericidal effect of lung and tracheal epithelial SPH. Upon loss of surface SPH, individuals become susceptible to P. aeruginosa infection, as exemplified in CF patients and mice, and in CerS2-null mice. Moreover, restoration of surface SPH by inhalation of SPH, of other LCBs or with AC, reverses susceptibility and cures existing P. aeruginosa infection, suggesting that LCB inhalation might provide a novel therapeutic option to counteract pulmonary infection by bacteria. This is of great importance since many P. aeruginosa strains are multi-resistant to antibiotics, rendering pulmonary infections difficult to treat. Acidification of the trachea, which increases AC activity, leads to a reduction in ceramide levels and an increase in sphingosine levels, consistent with the clinical use of hypertonic saline (pH 5.0-5.5) in treating CF patients (Elkins et al, 2006).
Ours is not the first study to identify SPH as an anti-bacterial agent (Bibel et al, 1992;Arikawa et al, 2002;Fischer et al, 2013). However, it is the first to show that LCBs are efficacious in acute lung infection. The mechanism of bacterial cell death may involve a direct effect of the LCBs on the bacteria (Fischer et al, 2012), or possibly up-regulation of porin-like proteins that are responsible for channel formation in the bacterial membrane (LaBauve & Wargo, 2014). Information obtained by comparing LCB structures suggests that longer and more positively charged LCBs are more effective than shorter, negatively charged LCBs; moreover, modification of the C-1 position of the LCB has variable effects on bacterial survival. Based on this structural information, we predict that it will be possible to generate LCBs that are highly efficacious in killing bacteria without causing effects on the host. If this notion is correct, then these putative novel LCBs analogs could pave the way for generation of drugs used for inhalation against P. aeruginosa infection and possibly infection by other bacteria.

Mice
CerS2-null and CF mice were generated as described (Charizopoulou et al, 2006;Pewzner-Jung et al, 2010). Two different CF mouse strains were used, Cftr tm1Unc -Tg (FABPCFTR) , abbreviated Cftr KO (mice that lack Cftr but express human CFTR in the gut under the control of a fatty acid binding protein promoter), obtained from Jackson Laboratories (Bar Harbor, ME, USA) and backcrossed for 10 generations onto a C57BL/6 background. All data shown in the current paper are from Cftr KO mice. To confirm data obtained with this strain, we used an additional strain, that is B6.129P2(CF/3)-Cftr TgH (neoim)Hgu (abbreviated Cftr MHH ) congenic mice that were established by brother-sister mating from the original Cftr TgH(neoim)Hgu mutant   mouse that was generated using insertional mutagenesis in Cftr exon 10. Congenic Cftr MHH mice were generated by backcrossing the targeted mutation onto the B6 inbred background. The strain produces low levels of Cftr. CerS2-null mice were on an F1 background achieved by intercross of heterozygous CerS2 GT/+ -C57BL/6 mice (GT = gene trap) and heterozygous CerS2 GT/+ -129S4/SvJae mice, to generate CerS2 GT/GT F1, and CerS2 +/+ F1 mice used in the experiments as CerS2-null and wild-type control, respectively. Control mice were syngenic littermates (C57BL/6 for CF mice and C57BL/6;129S4/SvJae WT for CerS2-null mice). The cystic fibrosis and the corresponding wild-type mice were used at an age of 16-18 weeks. We used female CF and WT mice in the present study. The CerS2-deficient mice were both male and females and used at an age of 6-8 weeks. Mice were bred in a special pathogen free (SPF) facility at the Weizmann Institute of Science and in the University of Duisburg-Essen. Mice were handled according to protocols approved by the Weizmann Institute of Science and the University of Duisburg-Essen Animal Care Committee as per international guidelines.
To exclude binding of the anti-SPH antibody to sphinganine, the trachea from WT or CF mice were removed, incubated in vitro with 10 lM sphinganine in 150 mM sodium acetate (pH 7.4) for 30 min, washed three times in 150 mM sodium acetate (pH 7.4), fixed for 36 h in 2% PFA (pH 7.3), and stained with the anti-SPH antibody. Hemalaun was used to stain lung sections for light microscopy.

Mouse inhalation
Inhalation was performed using a PARI Boy SX nebulizer (PARI GmbH, Starnberg, Germany), which generates a fine aerosol by pumping the fluid with an air jet. Mice were inhaled with the aerosol via a mask that is part of an oral inhalation device for children (LL-Nebulizer); the mask was clipped at the sides to cover only the nose and the surrounding part of the face. Mice were inhaled with 800 ll of 0.9% NaCl containing SPH (125 lM), AC (80 lg of purified protein), or FTY720 (125 lM). Approximately 10% of the volume that is applied to the mice is inhaled.

Enzymatic assays to measure ceramide and SPH analysis in situ
Mice were sacrificed and the trachea immediately removed. The trachea was carefully opened, washed in 150 mM sodium acetate (pH 7.4), placed on a 30°C pre-warmed plastic plate, and incubated with 0.01 units diacylglycerol (DAG) kinase (Biomol, Germany) to measure ceramide, or with 0.001 units SK (R&D Systems, Germany) to measure SPH, in 4 ll of 150 mM sodium acetate (titrated to pH 7.4), 1 mM adenosine triphosphate (ATP), and 10 lCi [ 32 P]cATP (for the buffer composition please see also Supplementary Fig S7). Controls were incubated with the same buffer without SK or left untreated. We ensured that the kinase buffer was only added to the luminal surface of the trachea. The kinase reaction for ceramide was performed for 15 min at 30°C and terminated by transfer of the trachea into CHCl 3 :CH 3 OH:1N HCl (100:100:1, v/v/v) followed by addition of 170 ll buffered saline solution (135 mM NaCl, 1.5 mM CaCl 2 , 0.5 mM MgCl 2 , 5.6 mM glucose, 10 mM HEPES, pH 7.2) and 30 ll of 100 mM EDTA. The SK reaction was terminated by placing the trachea in 100 ll H 2 O, followed by addition of 20 ll 1N HCl, 800 ll CHCl 3 /CH 3 OH/1N HCl (100:200:1, v/v/v), 240 ll CHCl 3 , and 2 M KCl. The lower phase was collected, dried, dissolved in 20 ll of CHCl 3 :CH 3 OH (1:1, v/v), and separated on Silica G60 thin layer chromatography (TLC) plates using CHCl 3 / acetone/CH 3 OH/acetic acid/H 2 O (50:20:15:10:5, v/v/v/v/v) as developing solvent for ceramide and CHCl 3 /CH 3 OH/acetic acid/ H 2 O (90:90:15:5, v/v/v/v) for SPH. The TLC plates were exposed to radiography films, spots were removed from the plates, and the incorporation of [ 32 P] into ceramide measured by liquid scintillation counting. Ceramide and SPH were determined using a standard curve of C16-ceramides to C24-ceramides or C18-SPH. Surface pH was varied by adjusting the pH of the 150 mM sodium acetate buffer from 7.4 to 5.0 and incubation of the trachea for 20 min prior to the in situ kinase assay. To exclude acid-mediated lysis of ceramide, we pre-incubated the trachea for 15 min with 0.5 mM oleoylethanolamine (Calbiochem) or 1 lM carmofur (Sigma) prior to adjusting the pH.

SPH analysis in extracts of freshly isolated tracheal epithelial cells
Epithelial cells were removed from the trachea by carefully scraping the inner surface of the trachea. Cells were extracted in CHCl 3 / CH 3 OH/1N HCl (100:200:1, v/v/v), and the lower phase was dried and resuspended in a detergent solution (7.5% [w/v] n-octyl glucopyranoside, 5 mM cardiolipin in 1 mM diethylenetriaminepentaacetic acid [DTPA]). The kinase reaction was initiated by addition of 0.001 units SK in 50 mM HEPES (pH 7.4), 250 mM NaCl, 30 mM MgCl 2 1 mM ATP, and 10 lCi [ 32 P]cATP. Samples were incubated for 30 min at 37°C with shaking (350 rpm) and processed as above.

Mass spectrometry
Ceramides and SPH in isolated tracheal epithelial cells were extracted and quantified as described (Fayyaz et al, 2014 The kinase reaction was terminated after 30 min, and samples extracted and processed as above.

Cytochalasin B treatment
Trachea was incubated with 10 lM cytochalasin B for 20 min prior to the kinase assay or prior to SPH immunoprecipitation experiments. Surface SPH was measured by the in situ kinase assay or by immunoprecipitation as above. As controls, trachea was incubated with cytochalasin B prior to infection with 1 × 10 6 CFU P. aeruginosa ATCC 27853. The trachea was washed after infection, incubated for 60 min with 100 lg/ml polymixin, washed again, lysed in 5 mg/ml saponin for 10 min, and centrifuged (10 min, 1,600 g). Pellets were resuspended in H/S buffer, aliquots were plated on TSA, and grown overnight prior to quantification of the number of intracellular bacteria.

AC activity assay
Mice were anesthetized with ketamine and xylazine and 4 ll   Cystic fibrosis (CF) is caused by mutations in the cystic fibrosis transmembrane conductance regulator and is the most common autosomal recessive disorder in western countries. At present, lung symptoms determine the quality of life and life expectancy of most CF patients, with patients displaying chronic inflammation and high susceptibility to lung infection with Pseudomonas aeruginosa, Haemophilus influenzae, Burkholderia cepacia, Staphylococcus aureus, and other bacteria. Approximately 80% of CF patients suffer from chronic P. aeruginosa pneumonia by the age of 25. Pulmonary Pseudomonas aeruginosa infections are also of major clinical importance in patients with chronic obstructive pulmonary disease, trauma, burn wounds, sepsis, or in patients requiring ventilation.

Results
Lung epithelial cells from human cystic fibrosis patients and cystic fibrosis mice display reduced sphingosine levels due to reduced acid ceramidase activity, which is reversed by acid ceramidase or sphingoid long chain base inhalation. Sphingoid long chain bases kill a broad spectrum of pathogenic bacteria at nanomolar to low micromolar concentrations, including Pseudomomas aeruginosa, Acinetobacter baumannii, Haemophilus influenzae and Moraxella catarrhalis and even Burkholderia cepacia. Inhalation of cystic fibrosis mice with acid ceramidase or sphingosine prevents and cures pulmonary Pseudomonas aeruginosa infection.

Impact
Tracheal and bronchial epithelial sphingosine acts as a natural antibacterial agent to prevent bacterial lung infection in healthy individuals. Sphingoid long chain bases show a broad anti-bacterial activity, and inhalation of sphingoid long chain bases may provide a novel therapeutic option to prevent or cure pulmonary bacterial infections.
designed the study and wrote the manuscript. All authors commented on the manuscript.