Apoptotic sphingolipid signaling by ceramides in lung endothelial cells

Abstract

The de novo pathway of ceramide synthesis has been implicated in the pathogenesis of excessive lung apoptosis and murine emphysema. Intracellular and paracellular-generated ceramides may trigger apoptosis and propagate the death signals to neighboring cells, respectively. In this study we compared the sphingolipid signaling pathways triggered by the paracellular- versus intracellular-generated ceramides as they induce lung endothelial cell apoptosis, a process important in emphysema development. Intermediate-chain length (C(8:0)) extracellular ceramides, used as a surrogate of paracellular ceramides, triggered caspase-3 activation in primary mouse lung endothelial cells, similar to TNF-alpha-generated endogenous ceramides. Inhibitory siRNA against serine palmitoyl transferase subunit 1 but not acid sphingomyelinase inhibited both C(8:0) ceramide- and TNF-alpha (plus cycloheximide)-induced apoptosis, consistent with the requirement for activation of the de novo pathway of sphingolipid synthesis. Tandem mass spectrometry analysis detected increases in both relative and absolute levels of C(16:0) ceramide in response to C(8:0) and TNF-alpha treatments. These results implicate the de novo pathway of ceramide synthesis in the apoptotic effects of both paracellular ceramides and TNF-alpha-stimulated intracellular ceramides in primary lung endothelial cells. The serine palmitoyl synthase-regulated ceramides synthesis may contribute to the amplification of pulmonary vascular injury induced by excessive ceramides.

Figure 1.

Kinetics of endogenous ceramide and caspase-3 up-regulation induced by exogenous ceramide (C_8:0 Cer) and TNF-α in primary lung endothelial cells. (a) Endogenous ceramides were measured by mass spectrometry, then normalized by lipid phosphorus (Pi). The data are presented as fold change compared with controls. In shaded circles are shown vehicle controls compared with untreated controls. When compared with controls, both exogenous C_8:0 ceramide (10 μM) and TNF-α (20 ng/ml with cycloheximide) increased endogenous ceramide levels (mean ± SEM; *P < 0.05 versus controls, Student's t test). (b) Caspase-3 activity (normalized by protein concentration and expressed as fold change compared with control) in endothelial cells treated with exogenous C_8:0 ceramide (10 μM) (mean ± SEM; *P < 0.05 versus controls, Student's t test). (c) Caspase-3 activity (normalized by protein concentration and expressed as fold change compared with control) in endothelial cells treated with TNF-α (20 ng/ml with cycloheximide) (mean ± SEM; *P < 0.05 versus controls, Student's t test).

Figure 2.

Effect of targeted siRNA treatment on SPT (serine palmitoyl transferase) and ASMase (acid sphingomyelinase) in lung endothelial cells. (a) Immunofluorescence micrograph of primary mouse lung endothelial cells treated with vehicle alone, siPortamine (left) and vehicle with fluorescently-labeled siRNA, SiGlo (right), showing a high efficiency (> 90%) of siRNA transfection (figures representative of n = 3). (b) SPT activity was measured as D₃-serine incorporation (% of total labeled serine administered to cells) by mass spectrometric analysis of labeled dihydrosphinganine and sphingosine (n = 3). Cells were assessed 6 days after treatment with SPTLC1 siRNA (50 nM). Inset: Upper immunoblot of SPT subunit 1 (57 kD) performed 6 days after treatment with SPTLC1 (SPT long-chain base subunit 1) siRNA at the indicated concentrations. Lanes: 1, untreated cell lysate; 2, siPortamine; 3, SPTLC1 siRNA, 50 nM; and 4, SPTLC1 siRNA, 100 nM. The lower immunoblot is for actin, shown as loading control. (c) Densitometry measurements of SPTLC immunoblots normalized by vinculin density as loading control of untreated cell lysates (CTL) and scramble siRNA (Scr SiRNA). Mean ± SEM (n = 3). (d) ASMase activity was measured fluorometrically at pH 5.0, as described in Materials and Methods, then normalized by protein concentration (n = 3). The experiments were performed in endothelial cells treated with the indicated concentrations of ASMase siRNA for 72 hours; (*P < 0.05, t test). Inset: representative (n = 3) immunoblot of ASMase (57 kD) 72 hours after treatment with ASMase siRNA at the indicated concentrations. Lanes: 1 and 2, untreated cell lysates; 3, Vehicle; 4, ASMase siRNA, 25 nM; and 5, ASMase siRNA, 50 nM. The same membrane was immunoblotted for GAPDH, as loading control. (e) Densitometry measurements of ASMase immunoblots normalized by vinculin density as loading control of untreated cell lysates (CTL) and scramble siRNA (Scr SiRNA). Mean ± SEM (n = 3).

Figure 3.

Effect of ceramide synthesis inhibition on effector caspase activation induced by extracellular ceramide in pulmonary endothelium. Primary mouse lung endothelial cells were pretreated with (a) SPTLC1 siRNA (50 nM, 6 d) or (b) ASMase siRNA (50 nM, 72 h) before C₈-ceramide treatment (10 μM, 18 h; Caspase-3/7 activity was normalized by protein concentration; Veh (vehicle siPortamine; mean + SD). Note the robust inhibition of C₈-ceramide–induced caspase activation by the specific inhibition of SPTLC1 compared with the vehicle control. (c) Effect of SPTLC1 siRNA on endogenous endothelial cell ceramide content (C_14:0 and higher fatty acid chain length) measured by mass spectrometry at 16 hours (mean ± SD, *P < 0.05 versus control cells; ^#P < 0.05 versus control C_8-ceramide–treated cells).

Figure 4.

Effect of ceramide synthesis inhibition on TNF-α–induced effector caspase activation in pulmonary endothelial cells. Primary mouse lung endothelial cells were pretreated with (a) SPTLC1 siRNA (50 nM, 6 d) or (b) ASMase siRNA (50 nM, 72 h) before TNF-α treatment (20 ng/ml, with cycloheximide). Caspase-3/7 activity was normalized by protein concentration; Veh (vehicle siPortamine). (c, d) Effect of siRNA inhibitory treatments on intracellular ceramides (normalized by lipid phosphorus). Effect of Scramble siRNA (Scr siRNA, 50 nM, 6 d) or SPTLC1 siRNA (50 nM, 6 d) on TNF-α-induced ceramides (c) compared with the effect of ASMase siRNA (50 nM, 3 d) (d). SPTLC1, but not ASMase inhibition significantly reduced ceramide up-regulation in response to TNF-α (20 ng/ml, with cycloheximide, 16 h); (mean ± SD, *P < 0.05 versus control untreated cells; ^#P < 0.05 versus control TNF-α–treated cells).

Figure 5.

Patterns of intracellular ceramide species induced by TNF-α or exogenous ceramides in lung endothelial cells. (a) Effect of TNF-α (20 ng/ml with cycloheximide, 16 h) or extracellular C_8:0 ceramide (10 μM, 16 h) treatment on endogenous long-chain (C_14:0 - C_18:0) and very long–chain ceramides (C_20–24). (b) Inhibitory effect of SPTLC1 and ASMase siRNAs on long-chain and very long–chain ceramides induced by TNF-α or C_8:0 ceramide. Note the robust inhibition of very-long-chain ceramides induction by SPTLC1 siRNA (n = 2). (c) Pie charts of relative expression profiles of individual ceramide species comprising total ceramides measured by mass spectrometry. Ceramide species expression profiles at baseline and in response to C_8:0 and TNF-α treatment. Depicted in the lower pie charts are the effects of SPTLC1 siRNA treatment on ceramide species induced by the two stimuli. (d) Changes in the relative abundance of C_16:0 (% of total ceramides) induced by C_8:0 and TNF-α treatment and the effect of SPTLC1 siRNA. The control values include untreated control, vehicle-treated control and scramble siRNA (mean ± SEM; *P < 0.05 versus controls, ^#P < 0.05 versus C_8:0 and TNF-α treatment, respectively, Student's t test).