Fetal calf serum (FCS) was purchased from HyClone (Logan, UT, USA) and TNF-α was from Genzyme (Cambridge, MA, USA). BAY 11–7085 (BAY) (purity ≥98%), medium 199, o-phenylenediamine and LPS from E. coli serotype 055:B5 were purchased from Sigma-Aldrich (Vienna, Austria). The 4,8-SD obtained after hydrolysis of GluCer was isolated from Arisaema amurense Maxim. (Araceae) and purified by silica column chromatography (CC). Polyclonal antibodies were purchased from R&D Systems (Minneapolis, MN, USA) and peroxidase-conjugated secondary antibodies from Amersham Life Science (Amersham, UK). For extraction, fractionation and isolation by CC solvents of highest available purity were used (VWR, Vienna, Austria). All other chemicals were obtained from Sigma-Aldrich.

Dried, processed rhizomes of Arisaema amurense Maxim. were purchased from Plantasia (Oberndorf, Austria). A voucher specimen, encoded ER-I, was deposited at the Department of Pharmacognosy, University of Vienna.

The dried, processed rhizomes of A. amurense (1.9 kg) were pulverized and extracted by dichloromethane and subsequently 1.4 kg of this material was extracted with methanol. The dried extracts were chromatographed over a silica column eluted by CHCl₃:MeOH:H₂O 98:2:1 to 60:38:8.5 to obtain fractions D1–8 for the dichloromethane extract (9.8 g) and eluted by CHCl₃:MeOH:H₂O 70:22:3.5 to 60:40:10 for the methanol extract (10.7 g) to obtain fractions M1–15. GluCer were isolated from fractions D8 (3.6 g) and M1 (2.5 g) by RP C-18 CC with methanol as mobile phase and subsequent silica CC with CHCl₃:MeOH:H₂O (65:25:4) as mobile phase (Fig. 1). A mixture of GluCer (40 mg) was hydrolyzed in methanolic 1 M HCl for 7 h at 70˚C under reflux. The hydrolyzed sample was neutralized and evaporated to dryness. The liberated sphin-goid base was purified by silica CC with CHCl₃:MeOH:NH₄ (40:10:1) as mobile phase (Fig. 2), with a yield of 5.5 mg.

The pure compound was further converted to the N-acetyl-di-O-trimethylsilyl derivative and analyzed by GC-MS and NMR and identified as 4,8-sphingadienine (4,8-SD) (Fig. 3). The amide group was N-acetylated with acetic anhydride:MeOH (v/v 1:4) overnight at 25˚C. The sample was brought to dryness and further derivatized by BSTFA and TMCS (v/v 99:1) and incubated at 50˚C for 30 min. The analyses were performed on an Agilent Technologies 6890N Network GC equipped with an Agilent Technologies 5973 inert Mass selective Detector and a CombiPAL autosampler (CTC Analytics). The column used was a DB-5 with dimensions of 30 m x 0.25 mm (inner diameter), 0.23 μm (film thickness) (Agilent Technologies). The software used was MSD Chemstation 2004.

NMR spectra were recorded on a Bruker Avance 500 NMR spectrometer. 4,8-SD was dissolved in CDCl₃ (99.96 atom % D). The ¹H and ¹³C NMR spectra were operated at 500 and 125 MHz, respectively. The compound has a purity of at least 95% judging from the NMR-spectra.

4,8-SD was a brownish, yellow, amorphous solid. Electron impact mass spectrometry (EIMS) of the N-acetyl-di-O-trimethylsilyl derivative showed ions at m/z 73, 174, 309, 334, 378, 468 (16). For biological testing 4,8-SD was dissolved in DMSO.

The study was performed using immortalized human umbilical vein endothelial (HUVECtert) cells (17). In contrast to primary HUVECs, HUVECtert have the advantage of an indefinite cell division potential. However, the tert transfection alters the cells’ gene expression profile (18) affective the response to the dual treatment. HUVECtert were cultured in M199 medium supplemented with endothelial cell growth supplement (Technoclone), heparin, penicillin/streptomycin and 20% FCS. The cells were grown in a humidified atmosphere at 37˚C and 5% CO₂ and passaged twice a week using trypsin-EDTA solution. Experiments were performed using cells up to passage 5.

HUVECtert cells (1x10⁵ cells/well) were seeded in a 96-well plate and grown for 24 h overnight in a cell culture incubator. Monolayers of the grown HUVECtert cells were then treated with 4,8-SD or GluCer and agonists in medium 199 containing 5% FCS. 4,8-SD (0–45 μM) or BAY 11–7082 (5 μM), which served as positive control, were added to the cells 20 min before the addition of TNF (100 ng/ml) or LPS (100 ng/ml). BAY 11-7085 is an anti-inflammatory agent acting by inhibiting NF-κB thus decreasing expression of inflammatory molecules including E-selectin and IL-8 (19). The solvent vehicle (0.2% DMSO) served as the negative control. After 6 h, the medium was collected for quantification of IL-8, and the cells were washed and fixed with 4% glutaraldehyde for determination of E-selectin. Cell surface-expressed E-selectin was detected using corresponding antibodies, secondary peroxidase-conjugated antibodies and o-phenylenediamine as substrate essentially as described before (20). Concentrations of IL-8 in cell culture medium were determined using human CXCL8/IL-8 ELISA DuoSet ELISA Development kit (R&D Systems) and the TMB 2-Component Microwell Peroxidase Substrate kit (VWR International, Radnor, PA, USA). The absorbance was measured at 490 nm.

The HUVECtert cells (1x10⁵ cells/well) were seeded in 96-well plates. After 24 h test compounds (4,8-SD or GluCer) or vehicle (DMSO) were added. Cell viability and metabolic activity were assessed after 6 h either by the LDH (Sigma Aldrich, St. Louis, MO, USA) or WST-1 (Roche Applied Science, Mannheim, Germany) colorimetric assays. The absorbance was measured at 340 nm after 55 min for the LDH assay at 450 nm for the WST1-assay.

Statistical analysis and calculation of IC₅₀ values were performed using the Prism Software (ver. 4.03; GraphPad Software Inc., San Diego, CA, USA). Data were normalized to the mean value from three experiments of DMSO treated control. The data shown represent the mean values out of at least three experiments ± SEM. Statistical significance was determined by a one-way analysis of variance combined with a Dunnett’s multiple comparison post test. Results with P<0.05 were considered significant.