Platelet-derived exosomes induce endothelial cell apoptosis through peroxynitrite generation: experimental evidence for a novel mechanism of septic vascular dysfunction

Abstract

Introduction: Several studies link hematological dysfunction to severity of sepsis. Previously we showed that platelet-derived microparticles from septic patients induce vascular cell apoptosis through the NADPH oxidase-dependent release of superoxide. We sought to further characterize the microparticle-dependent vascular injury pathway.

Methods: During septic shock there is increased generation of thrombin, TNF-alpha and nitric oxide (NO). Human platelets were exposed for 1 hour to the NO donor diethylamine-NONOate (0.5 microM), lipopolysaccharide (LPS; 100 ng/ml), TNF-alpha (40 ng/ml), or thrombin (5 IU/ml). Microparticles were recovered through filtration and ultracentrifugation and analyzed by electron microscopy, flow cytometry or Western blotting for protein identification. Redox activity was characterized by lucigenin (5 microM) or coelenterazine (5 microM) luminescence and by 4,5-diaminofluorescein (10 mM) and 2',7'-dichlorofluorescein (10 mM) fluorescence. Endothelial cell apoptosis was detected by phosphatidylserine exposure and by measurement of caspase-3 activity with an enzyme-linked immunoassay.

Results: Size, morphology, high exposure of the tetraspanins CD9, CD63, and CD81, together with low phosphatidylserine, showed that platelets exposed to NONOate and LPS, but not to TNF-alpha or thrombin, generate microparticles similar to those recovered from septic patients, and characterize them as exosomes. Luminescence and fluorescence studies, and the use of specific inhibitors, revealed concomitant superoxide and NO generation. Western blots showed the presence of NO synthase II (but not isoforms I or III) and of the NADPH oxidase subunits p22phox, protein disulfide isomerase and Nox. Endothelial cells exposed to the exosomes underwent apoptosis and caspase-3 activation, which were inhibited by NO synthase inhibitors or by a superoxide dismutase mimetic and totally blocked by urate (1 mM), suggesting a role for the peroxynitrite radical. None of these redox properties and proapoptotic effects was evident in microparticles recovered from platelets exposed to thrombin or TNF-alpha.

Conclusion: We showed that, in sepsis, NO and bacterial elements are responsible for type-specific platelet-derived exosome generation. Those exosomes have an active role in vascular signaling as redox-active particles that can induce endothelial cell caspase-3 activation and apoptosis by generating superoxide, NO and peroxynitrite. Thus, exosomes must be considered for further developments in understanding and treating vascular dysfunction in sepsis.

Figures

Figure 1

Tetraspan protein enrichment characterizes exosomes. The graph shows the percentage of positive events per 100,000 counts as analyzed by flow cytometry. Values are corrected for background and non-specific antibody binding. Exosomes obtained from septic patients as well as from platelets activated by the nitric oxide donor diethylamine-NONOate (NONOate; 0.5 μM) or lipopolysaccharide (LPS; 100 ng/ml) expose larger amounts of tetraspan protein family members CD9, CD63, and CD81, and less phosphatidylserine (as assessed by annexin V staining) than particles obtained from platelets treated only with saline (Control) or thrombin (5 IU/ml) or from apoptotic endothelial cells (apoptosis). Results are means ± SD. For each bar, n = 4 samples. *P < 0.05 versus control, †P < 0.05 versus apoptotic bodies (apoptosis). UV, ultraviolet.

Figure 2

Electron microscopy reveals the structure of exosomes. Images obtained from the exosome population generated by platelets exposed to diethylamine-NONOate (a) and to thrombin (b) reveal rounded membranaceous structures measuring on average less than 150 nm. It is noteworthy that exosomes from platelets stimulated with diethylamine-NONOate have a more regular surface than those generated by platelets exposed to thrombin. Scale bars, 100 nm; original magnification ×60,000.

Figure 3

Lucigenin luminescence: exosomes from platelets exposed to NO or LPS are similar to septic exosomes. The graph represents NADPH-dependent lucigenin (5 μM) chemiluminescence above background. Exosomes (10 μg protein content) obtained from platelets exposed to the nitric oxide donor diethylamine NONOate (NONOate; 0.5 μM) or lipopolysaccharide (LPS; 100 ng/ml) generate reactive oxygen species in a similar fashion to exosomes obtained from septic patients, whereas particles obtained from platelets exposed to saline (control) or thrombin (5 IU/ml) have very low activity. For comparison, luminescence obtained with platelets from healthy (control) and septic subjects are displayed. Results normalized for sample protein concentration are means ± SD of three or more experiments. *P < 0.05 versus control.

Figure 4

Coelenterazine luminescence triggered by exosomes suggests the presence of reactive oxygen and nitrogen generation. The graph represents exosome coelenterazine (5 μM) luminescence above background. Exosomes were incubated with NADPH and L-arginine. Exosomes (10 μg protein content) obtained from platelets exposed to the nitric oxide donor diethylamine NONOate (NONOate; 0.5 μM) or lipopolysaccharide generate reactive oxygen species in a similar fashion to exosomes obtained from septic patients, whereas particles obtained from platelets exposed to saline (control) or thrombin have very low activity. Luminescent signals were consistently inhibited by the addition of the superoxide dismutase mimetic Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (SOD, 10 μM) and by the NO synthase inhibitors L-NMA (NG-methyl-L-arginine acetate; 5 mM), or Nω-Nitro-L-arginine methyl ester (L-NAME; 1 mM), suggesting the generation of reactive oxygen species and reactive nitrogen species by the exosomes. Results are means ± SD of seven experiments. *P < 0.05 versus control, †P < 0.05 versus untreated. RLU, relative luminescence units.

Figure 5

NADPH oxidase and uncoupled NO synthase are sources of reactive species from platelet-derived exosomes. Exosomes from septic patients, as well as exosomes induced with the nitric oxide donor diethylamine NONOate (NONOate; 0.5 μM) and lipopolysaccharide (LPS) caused enhanced 2',7'-dichlorofluorescein diacetate (10 mM) fluorescence (after the addition of 100 μM NADPH), which was significantly inhibited by the membrane-permeable superoxide dismutase mimetic Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (SOD) or by the NADPH oxidase-blocking peptide gp91 ds-tat (10 μM), confirming the role of a superoxide-generating NADPH oxidase. Nω-nitro-D-arginine methyl ester (L-NAME) decreased the fluorescent signals, suggesting a role for uncoupled nitric oxide synthase in superoxide generation. The scrambled peptide (scr ds-tat) used as a control for gp91 ds-tat shows a non-significant residual inhibitory effect. Results are means ± SD of five experiments for each group. *P < 0.05 versus control, †P < 0.05 versus untreated. RFU, relative fluorescence units.

Figure 6

Platelet-derived exosomes may generate peroxynitrite. The graph shows a decrease in 2',7'-dichlorofluorescein diacetate signals after the addition of L-arginine (1 mM), further suggesting a role for uncoupled nitric oxide synthase in superoxide generation. In contrast, the inhibitory effect of urate addition strongly suggests the involvement of peroxynitrite oxidation. Results are means ± SD of five experiments for each group. *P < 0.05 versus control, †P < 0.05 versus untreated. NONOate, diethylamine NONOate; RFU, relative fluorescence units.

Figure 7

Exosomes generate reactive nitrogen species. The graphic shows 4,5-diaminofluorescein diacetate (10 mM) fluorescence of exosomes incubated with L-arginine. The membrane-permeable superoxide dismutase mimetic Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (SOD) had no inhibitory effect, whereas Nω-nitro-D-arginine methyl ester (L-NAME) and urate caused a significant decrease in fluorescent signals, suggesting the generation of reactive nitrogen species by exosomes, more importantly by septic exosomes and by exosomes induced with nitric oxide or lipopolysaccharide (LPS). Results are means ± SD of four experiments for each group. *P < 0.05 versus control, †P < 0.05 versus untreated. RFU, relative fluorescence units.

Figure 8

Platelet-derived exosomes possess NADPH oxidase and nitric oxide synthases. Representative Western blot images of exosomes from different origins (septic platelets (sepsis), platelets exposed to diethylamine-NONOate (NONO), lipopolysaccharide (LPS), TNF-α, thrombin (Thr) and saline (Ctl)) were subjected to SDS-PAGE and exposed to antibodies directed to the different nitric oxide synthase (NOS) isoforms: neuronal (nNOS), inducible (iNOS) and endothelial (eNOS), to the NADPH oxidase regulatory protein protein disulfide isomerase (PDI), to the NADPH oxidase membrane-bound subunit isoforms Nox 1 and Nox2, and to the NADPH oxidase membrane component p22phox. Leukocytes were used as positive controls for iNOS, and NADPH oxidase components, endothelial cells activated (+) or not (-) with LPS were used as controls for eNOS/iNOS expression. Results shown are representative of at least three different experiments.

Figure 9

Nitric oxide-induced and septic platelet-derived exosomes cause ROS/RNS-dependent apoptosis in endothelial cells. Exosomes obtained from septic patients or from platelets exposed to a nitric oxide (NO) donor (diethylamine-NONOate; NONOate) cause a twofold to threefold increase in apoptosis rates of rabbit endothelial cells compared with exosomes from platelets exposed to saline (not shown) or thrombin. The membrane-permeable superoxide dismutase mimetic Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (SOD; 10 mM), the NO synthase inhibitor Nω-nitro-L-arginine methyl ester (L-NAME; 1 mM), or the peroxynitrite scavanger urate (1 mM) reversed the proapoptotic activity of exosomes. Results are means ± SD of six experiments for each group. *P < 0.05 versus control, †P < 0.05 versus untreated. ROS, reactive oxygen species; RNS, reactive nitrogen species.

Figure 10

Exosomes cause reactive oxygen species/reactive nitrogen species-dependent caspase-3 activation in endothelial cells. Exosomes obtained from platelets exposed to saline (not shown) or thrombin did not cause caspase-3 activation above baseline in rabbit endothelial cells. In contrast, exosomes from septic patients (sepsis) or from platelets exposed to lipopolysaccharide (LPS) or a nitric oxide donor (diethylamine-NONOate; NONOate) caused a doubling of caspase-3 activation over baseline, similar to the activation obtained by direct exposure of endothelial cells to 40 ng/ml TNF-α (+TNF-α). The membrane-permeable superoxide dismutase mimetic Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (SOD) and Nω-nitro-L-arginine methyl ester (L-NAME) completely blocked exosome-triggered caspase-3 activation. Results are means ± SD of three experiments for each group. *P < 0.05 versus control, †P < 0.05 versus untreated.