This study was performed with the approval of and in strict accordance with the UK Home Office and the University of Liverpool ethics committee (Home Office Project License Number 40/3602) as well as the Bezirksregierung Düsseldorf and the University of Duisburg-Essen ethics committee (protocol number Az. 84-02.04.2013.A282) or the University of Cincinnati ethics committee (IACUC# 10-05-10-01). Every effort was made to minimize suffering, and in bacterial infection experiments, mice were humanely euthanized if they became lethargic. All animal experiments were carried out at the Universities of Liverpool, Duisburg-Essen and Cincinnati.

Reagents.

Streptolysin O from S. pyogenes, α-hemolysin from S. aureus, tetanolysin from Clostridium tetani, phospholipase C from Clostridium perfringens were from Sigma. Pneumolysin was prepared as previously described33. HDL (3.2 mg/ml cholesterol) was from Calbiochem. LPS from E. coli or P. aeruginosa was from Sigma. Deuterated sphingomyelin (16:0-d31) was from Avanti Polar Lipids.

Cells.

The human embryonic kidney cell line (HEK 293; ATCC CRL-1573) was maintained as previously described34. Human umbilical vein endothelial cells (HUVEC) were obtained from Promocell and maintained in Promocell's Endothelial Cell Growth Medium. The human acute monocytic leukemia cell line (THP-1; ATCC TIB-202) was maintained in RPMI 1640 medium containing 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin.

Bacterial cultures and supernatants.

Bacterial culture supernatants were prepared from S. pyogenes strains 50362, 31009 (clinical isolates from blood culture), 30979 (clinical isolate from tissue biopsy), 31116 (clinical isolate from nasopharynx) and DSM 11728; S. aureus strains MRSA 2040, MRSA 3208.21 (clinical isolate from nasopharynx), MRSA 3108.46 (clinical isolate from a wound) and a methicillin-sensitive S. aureus strain isolated from a septic patient (septic S. aureus strain) as well as from S. pneumoniae (clinical isolate, 111.46). The septic S. aureus strain, also used for in vivo experiments, produced enterotoxin D and was tested negative for Panton-Valentin-leukocidin (PVL) and toxic shock syndrome toxin (TSST).

The bacteria were grown overnight on blood agar plates, resuspended at an OD of 0.225 in 40 ml tryptic soy broth (TSB) and allowed to grow at 37 °C with shaking at 125 r.p.m. to an OD 600 = 0.6–0.8. Bacterial cultures were centrifuged at 2,800 r.p.m. for 10 min at room temperature. Culture supernatants were harvested and stored at −70 °C.

Liposomes.

Cholesterol, sphingomyelin from egg yolk, phosphatidylcholine from egg yolk and phospatidylserine (sodium salt) from bovine brain were from Sigma. The lipids were individually dissolved in chloroform at 1 mg/ml concentrations. For the preparation of liposomes the chloroform solutions of the individual lipids were mixed in the composition and the proportions, which are given in the text. Chloroform was completely evaporated for 30 min at 60 °C, followed by 2 h under vacuum. Tyrode's buffer (140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 10 mM glucose, 10 mM HEPES; pH = 7.4) containing 2.5 mM CaCl 2 was added to the films of dried lipids. Following incubation for 30 min at 45 °C in an Eppendorf thermomixer with vigorous shaking, the liposomes were sonicated three times for 5 s at 6 °C in a Bandelin Sonopuls sonicator at 80% power. The liposomal preparations were left for at least 1 h at 6 °C before they were used in experiments.

Alternatively, unilamellar cholesterol:sphingomyelin (66 mol/% cholesterol; mean diameter = 130 nm) and sphingomyelin (100%; mean diameter = 90 nm) liposomes, produced according to a good manufacturing practice (GMP)-compliant process protocol, were provided by Lascco (Geneva, Switzerland – product name CAL02). The amounts of liposomes are given as the amount of total lipids used for their preparation.

Toxin-induced cell lysis and protective effects of liposomes in vitro.

The effects of purified toxins or bacterial culture supernatants on the proliferation of a human monocyte cell line (THP-1) were assessed in the presence or absence of liposomes of various lipid compositions. Routinely, 100–200 μl of a purified toxin solution or a bacterial culture supernatant (Ca2+-Tyrode's buffer or bacterial culture broth were used as controls) were added to 100 μl (5 × 104 cells/ml or 5 × 105 cells/ml) of cells pre-mixed with 50 μl of liposomes. After incubation for 3 h, 2 ml of cell culture medium was added to the tubes. The cells were counted for 5–14 d. The toxins and the liposomes were present for the whole duration of an experiment.

In human embryonic kidney epithelial cells (HEK 293) and in human umbilical vein endothelial cells (HUVEC), toxin-induced lysis was monitored as a decline of cytoplasmic fluorescence due to an efflux of intracellular GFP (green-fluorescent protein) or YFP (yellow-fluorescent protein). GFP or YFP were transiently expressed as described34. GFP- or YFP-expressing cells were used for laser scanning module (LSM) imaging experiments 2 d after transfection. Cells, seeded on 15 mm glass coverslips (5 × 105 cells per coverslip), were mounted in a perfusion chamber at 25 °C in 200 μl of Tyrode's buffer containing 2.5 mM CaCl 2 , and their fluorescence was recorded in an Axiovert 200 M microscope with a laser scanning module LSM 510 META (Zeiss) using a × 63 oil immersion lens. At time-point 0, the buffer was replaced by 100 μl of solution containing liposomes, followed (with 20–30 s of handling delay) by 100 μl of the toxin-containing solution. The images were analyzed using the “Physiology evaluation” software package (Zeiss).

Pneumolysin binding to Ch:Sm liposomes.

Pneumolysin was diluted in Ca2+-Tyrode's buffer and centrifuged at 100,000g for 1 h to remove aggregated protein.

In a direct binding assay, 200 ng of pneumolysin was incubated with 0–9 μg of cholesterol:sphingomyelin liposomes (66 mol/% cholesterol) for 20 min at 20 °C. Liposomes were pelleted by centrifugation at 100,000g for 2 h. pneumolysin content in the pellets and in the supernatants was assessed by western blot analysis.

In a competition binding assay, 100 ng of pneumolysin was incubated with or without THP-1 cells (5 × 105 cells/ml) and with or without 100 μg of cholesterol:sphingomyelin liposomes (66 mol/% cholesterol) for 20 min at 20 °C. The samples were centrifuged at 134 g; supernatants were collected and further centrifuged at 100,000g for 2 h. Low-speed and high-speed pellets were dissolved in 50 μl of SDS-PAGE sample buffer. The protein of the high-speed supernatants was precipitated by 10% trichloroacetic acid in the presence of 1 mg/ml bovine serum albumin, pelleted and dissolved in 50 μl of SDS-PAGE sample buffer. Pneumolysin content in the samples was assessed by western blot analysis. Monoclonal antibody against pneumolysin (1F11; Santa Cruz Biotechnology) was used at 1:500 dilution.

Release of CXCL8 was measured in culture supernatants of HUVEC cells 24 h after treatment with 15 ng (30 ng/ml) of pneumolysin in the presence or absence of 40 μg/ml cholesterol:sphingomyelin (66 mol/% cholesterol) liposomes by Bioplex assay according to the manufacturer's instructions.

Protective effects of liposomes in vivo.

6- to 8-week-old, age matched, C57BL6/J male mice were used for S. aureus in vivo infection experiments. All experiments undertaken were approved by local authorities. S. aureus were grown overnight on blood agar plates, the bacteria were resuspended at an OD of 0.225 in 40 ml TSB and allowed to grow at 37 °C with shaking at 125 r.p.m. to the early logarithmic phase for 1 h. Bacterial cultures were then pelleted by centrifugation at 2,800g r.p.m. and washed in phosphate buffered saline. 3 × 106 CFU bacteria were resuspended in 100 μl of 0.9% saline and injected intravenously into mice. A 1:1 mixture of cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes was injected at the indicated times and doses in 0.9% saline. Controls received saline only. Survival was determined over 7 d.

6- to 8-week-old, age matched, MF1 (Charles River, UK) female mice were used for S. pneumoniae infection experiments. Bacterial inoculum for infection was prepared from stocks frozen in mid-log phase. Aliquots were thawed, centrifuged (13,000g, 2 min) and washed with endotoxin-free PBS before infection. For induction of pneumococcal pneumonia, mice were anesthetized with a mixture of O 2 and isofluorane and infected intranasally with 1 × 106 CFU S. pneumoniae serotype 2 (strain D39) in PBS. Liposome preparations were administered intranasally 30 min after infection. For induction of bacteremia, 1 × 106 CFU D39 were injected into the tail vein. Liposome preparations were administered intravenously at the indicated times and doses. Tail bleeds were done 24 h after infection to assess bacteremia.

In all experiments mice were monitored for signs of disease using the scheme of Morton35 and were euthanized once disease had reached the “lethargic” state.

Biodistribution of liposomes.

Cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes, each containing 6% of deuterated sphingomyelin were prepared as described above. Healthy mice and mice that were intravenously infected with S. aureus were injected with 100 mg/kg of the deuterated liposomes. Mice were euthanized immediately, 6.5 h or 24 h after liposomal injection (n = 4 for each time point). Blood, lungs, kidneys, aortas, livers, intestines and spleens were removed and analyzed by mass spectrometry. 16:0-d31 sphingomyelin was extracted and quantified as recently described36. Sample analysis was carried out by rapid resolution liquid chromatography-MS/MS using a Q-TOF 6530 mass spectrometer (Agilent Technologies) operating in the positive electrospray ionization mode. Chromatographic separation on the Zobrax Eclipse column (C8, 2.1 × 150 mm, 3.5 μm particle size; Agilent Technologies) was achieved by gradient elution using mobile-phase A (water/formic acid 100:0.1 v/v) and mobile-phase B (acetonitrile/methanol/formic acid 50:50:0.1 v/v). The precursor ions of 12:0 sphingomyelin (m/z 647.5123) as internal standard and 16:0-d31 sphingomyelin (m/z 734.7619) were cleaved into the fragment ion of m/z 184.3. Quantification was performed with Mass Hunter Software (Agilent Technologies).

Antibiotic treatment.

In the pneumococcal sepsis model, penicillin (25 mg/kg) was given intraperitoneally (i.p.) at 10 and 24 h after infection. In the staphylococcal sepsis model, vancomycin (100 mg/kg) was given i.p. at 1 and 9 h after infection.

Injection of bacterial supernatants.

S. aureus was grown overnight on blood agar plates. Bacteria were resuspended at an OD of 0.225 in 40 ml TSB and allowed to grow at 37 °C with shaking at 125 r.p.m. to the early logarithmic phase for 1 h. Bacteria were centrifuged at 3,200 r.p.m., the supernatant collected, centrifuged again, a 150 μl aliquot of the supernatant collected and incubated for 4 h with 100 mg/kg liposomes consisting of a 1:1 mixture of cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes at 37 °C. Controls were supplemented with the same volume PBS. After the 4 h incubation, samples were centrifuged at 20,000g and 150 μl of supernatant was injected into mice i.v.

Liposomal treatment of LPS-septic mice.

LPS from E. coli or P. aeruginosa was injected i.p. 45 min before intravenous application of liposomes.

TNF-alpha release.

TNF-α was measured in lung homogenates or serum (diluted 1:10 in PBS) obtained from healthy mice; healthy liposome-treated mice; mice infected with S. pneumoniae D39; and D39-infected mice that were treated with the liposomes by enzyme-linked immunosorbent assay (ELISA) (BioLegend), according to manufacturer's instructions.

Blood leukocytes.

Blood leukocytes were counted in lung homogenates or serum (diluted 1:10 in PBS) by flow cytometry. Mouse tissue cell suspensions, at 2 × 108 cells/ml were incubated with purified anti-Fc receptor blocking antibody (anti-CD16/CD32) before addition of antibodies against cell surface markers. A combination of fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, PE-Cy7- and allophycocyanin (APC)-conjugated monoclonal antibodies (eBioscience) were used. CD19 was used to identify B cells, CD3, CD4 and CD8 to identify T cells and Gr-1 and CD11b to identify neutrophils. In each experiment the appropriate monoclonal control antibodies and single conjugate controls were included. Alternatively, blood leukocytes were quantified by their typical forward versus sideward scatter using flow cytometry. Samples were analyzed using a Becton Dickinson FACScalibur flow cytometer running CellQuest acquisition and analyzed using FlowJo software (Tree Star).

Lung edema in vivo after S. aureus infection.

S. aureus was grown overnight on blood agar plates, the bacteria were resuspended at an OD of 0.225 in 40 ml TSB and allowed to grow at 37 °C with shaking at 125 r.p.m. to the early logarithmic phase for 1 h. Bacteria were centrifuged at 2,800 r.p.m., washed once in PBS and 5 × 106 CFU were i.v. injected in a volume of 100 μl. Controls were injected with PBS. We then determined survival and lung edema by measuring wet and dry weight as well as Evans Blue leakage 12 h after the infection. Liposomes (100 mg/kg; 1:1 mixture of cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes) were injected 2 and 6 h after the infection. Evans Blue (20 mg/kg) was i.v. injected, mice were euthanized after 10 min, the lung was flushed with PBS via the right heart, removed and dried for 2 d at 56 °C. The dried tissue was suspended in 0.8 ml formamide per 100 mg dry tissue and incubated for 24 h at 60 °C. The OD of the samples was then analyzed by 620 nm and the concentration determined using a standard curve of Evans Blue.

Killing capacity of blood samples.

40 μl of mouse blood were incubated with 10,000 CFU S. aureus at 37 °C for 4 h with shaking at 125 r.p.m. in the presence of either liposomes at 1 mg/ml (1:1 mixture of cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes) or the same volume PBS. After the incubation, aliquots were plated on LB plates, grown o/n and the CFU were counted.

NO release.

Blood samples were taken from healthy mice and we performed a Percoll density gradient centrifugation to purify neutrophils. Neutrophils were incubated ex vivo with 5 μM 4-amino-5-methylamino-2′,7′-difluorofluorescein (DAFFM) (Molecular Probes) at 37 °C for 30 min. DAFFM is nonfluorescent and reacts with NO to a fluorescent substrate. Samples were washed with PBS and then further incubated with 0.5 mg/ml of each cholesterol:sphingomyelin (66 mol/% cholesterol) and sphingomyelin-only liposomes (1 mg/ml total) at 37 °C for 30 min or with the same volume of PBS. They were then challenged with S. aureus at an multiplicity of infection (MOI) of 1 cell per 10 bacteria and NO was determined over a period of 90 min in the presence or absence of L-NAME, an i-NOS inhibitor. Samples were then analyzed by fluorescence-activated cell sorting (FACS). The relative NO production induced by S. aureus, as is indicated by the change of the fluorescence of the DAFFM, was calculated by using the formula ΔDAFFM mean fluorescence after S. aureus = mean fluorescence after S. aureus – mean fluorescence without infection. Fluorescence intensity of the L-NAME group was calculated as following: ΔDAFFM mean fluorescence after S. aureus + L-NAME = mean fluorescence after S. aureus + L-NAME − mean fluorescence L-NAME only.

ROS release.

Blood samples were taken from healthy mice, and we performed a Percoll density gradient centrifugation to obtain neutrophils. The cells were incubated ex vivo with 10 μM 2′,7′ dichlorodihydrofluorescein diacetate acetyl ester (DCF) (H2DCFDA, Molecular Probes) at 37 °C for 30 min, incubated with 0.5 mg/ml of each liposome (1 mg/ml total as above) at 37 °C for 30 min or with the same volume of PBS, challenged with S. aureus at an MOI of 1 cell per 10 bacteria and ROS was determined over a period of 60 min. Controls were incubated with 100 U/ml superoxide dismutase (SOD) and 10 U/ml catalase. Samples were analyzed by FACS. The relative ROS production induced by S. aureus, as is indicated by the change of the fluorescence of the DCF, was calculated by using the formula ΔDCF mean fluorescence = mean fluorescence after S. aureus - mean fluorescence after [S. aureus + SOD + catalase].

Bactericidal activity of liposomes.

To define the minimum bactericidal concentration (MBC), we followed the guidelines proposed by the Clinical and Laboratory Standards Institute37. For the broth microtiter dilution tests, 96-well plates supplemented with 50 μl of 0.5–16 mg/ml liposomes were inoculated with 50 μl of Mueller Hinton broth containing a bacterial cell suspension of 1–5 × 105 colony-forming units (CFU)/ml. The plates were incubated for 24 h at 36 °C. MBC was tested by transferring 10 μl aliquots from the wells of the broth microtiter dilution plates onto Columbia blood agar (Oxoid). The inoculated plates were further incubated for 24 h at 36 °C and then colonies were counted.

Statistical analysis, sample size planning and measures against bias.

For the comparison of continuous variables from independent groups we used Student's t-test for two groups and one-way ANOVA for more than two groups followed by post-hoc Student's t-tests for all pairwise comparisons applying Bonferroni correction for multiple testing. Between-group comparisons of survival variables was performed by log-rank test.

As sensitivity analyses, which generally supported the claim of the prespecified analyses, we applied Welch's t-tests and exact Wilcoxon-Mann-Whitney tests for continuous variables and the latter also for survival variables given that no censoring events happened during the follow-up. Furthermore, to follow our sample size plan, we also applied Fisher's exact test to compare the survival rates at the last observation between groups. All reported P values are two-sided and corrected for comparison-wise multiplicity.

The sample size planning of the mouse experiments with outcome death was based on two-sided Fisher's exact tests (i.e., a comparison of event rates as suggested by Shah38) even though the pre-planned analyses was for survival times. When aiming to detect a between-group difference in survival rates between 15% and 90% (at the last follow-up, assuming no censoring events before), ten animals per group (1:1 allocation) are required for a comparison-wise power of 90% at a significance level α = 5% (two-sided); for n = 8/n = 12 per group the power will be about 87%/98%. The sample size planning for the continuous variables in, for example, the S. aureus in vivo infection experiments, was based on two-sided Wilcoxon-Mann-Whitney tests (software: G*Power Version 3.1.7 of the University of Duesseldorf, Germany). When aiming at detecting a large standardized effect of 2.5 (in units of a s.d. of a standard normal distribution) between means of two groups, four animals per group (1:1 allocations) are required for a comparison-wise power of 80% at a significance level α = 5% (two-sided). Thus, the animal experiments were powered to detect large effects.

Mice were divided into cages of equal size (usually three to five mice) on arrival by animal unit technical staff with no involvement in study design. Thus, for a single experiment, all mice used were of the same age and sex and were within 2 g weight of each other. Cages were numbered and the numbers randomly assigned to an experimental treatment group. The investigators were not blinded to the group allocation during the experiment and/or when assessing the outcome.

For the pain score, mice were scored separately by two independent researchers and one animal unit technical staff member with no involvement in study design and recorded.

Investigators were blinded for histology and Evans Blue experiments.