Endothelial transcytosis of myeloperoxidase confers specificity to vascular ECM proteins as targets of tyrosine nitration

Abstract

Nitrotyrosine formation is a hallmark of vascular inflammation, with polymorphonuclear neutrophil-derived (PMN-derived) and monocyte-derived myeloperoxidase (MPO) being shown to catalyze this posttranslational protein modification via oxidation of nitrite (NO(2)(-)) to nitrogen dioxide (NO(2)(*)). Herein, we show that MPO concentrates in the subendothelial matrix of vascular tissues by a transcytotic mechanism and serves as a catalyst of ECM protein tyrosine nitration. Purified MPO and MPO released by intraluminal degranulation of activated human PMNs avidly bound to aortic endothelial cell glycosaminoglycans in both cell monolayer and isolated vessel models. Cell-bound MPO rapidly transcytosed intact endothelium and colocalized abluminally with the ECM protein fibronectin. In the presence of the substrates hydrogen peroxide (H(2)O(2)) and NO(2)(-), cell and vessel wall-associated MPO catalyzed nitration of ECM protein tyrosine residues, with fibronectin identified as a major target protein. Both heparin and the low-molecular weight heparin enoxaparin significantly inhibited MPO binding and protein nitrotyrosine (NO(2)Tyr) formation in both cultured endothelial cells and rat aortic tissues. MPO(-/-) mice treated with intraperitoneal zymosan had lower hepatic NO(2)Tyr/tyrosine ratios than did zymosan-treated wild-type mice. These data indicate that MPO significantly contributes to NO(2)Tyr formation in vivo. Moreover, transcytosis of MPO, occurring independently of leukocyte emigration, confers specificity to nitration of vascular matrix proteins.

Figure 1

Immunohistochemical detection of aortic endothelial distribution of MPO following activation of PMNs. Isolated human PMNs (8 × 10⁵) were instilled into the lumen of an excised rat aorta immediately after activation with 50 ng·ml^–1 TPA. After 4 hours, aortic segments were frozen, sectioned, fixed with 4% paraformaldehyde, permeabilized using 0.1% Triton X-100, and analyzed by immunofluorescence. Tissue sections were stained for MPO (green) using rabbit anti-MPO and nuclei (blue) were counterstained with DAPI. (a) Untreated vessel; (b) vessel treated with TPA alone; (c) unstimulated PMNs; (d and e) stimulated PMNs. Colocalization of MPO with anti-vWF immunoreactivity confirmed MPO binding and uptake by endothelial cells (f). L, vessel lumen; M, media. Arrows indicate endothelial cells. ×100 (a–d, f); ×50 (e).

Figure 2

Characterization of the binding and uptake of MPO by cultured endothelial cells. (a) Characterization of cell MPO association. BAECs were incubated with increasing concentrations of MPO (2–13 nM) for 2 hours at 37°C and harvested by scraping (total) or after exposure to trypsin (trypsin-resistant). Cell-associated MPO activity was about 10% of total MPO activity added (not shown). (b) The influence of temperature on cellular MPO binding and transport. BAECs exposed to MPO (13 nM) at 37°C and 4°C were harvested at different time points and MPO activity was determined. ^†P < 0.05 for MPO activity in total cell lysates at 37°C versus 4°C; *P < 0.05 for MPO activity in trypsin-resistant compartments at 37°C versus 4°C. (c) The influence of glycosaminoglycans on cell MPO binding. BAECs were pretreated with chondroitin sulfate (Chondr, 150 μg·ml^–1), heparin (Hep, 150 μg·ml^–1), or the low–molecular weight heparin enoxaparin (Enox, 150 μg·ml^–1) for 45 minutes, washed with HBSS, and then exposed to MPO (13 nM) at 37°C. After washing again, cell-associated MPO was assessed by enzyme activity analysis and cellular MPO protein content by immunoblotting. “MPO hc ” denotes the immunoreactivity of the heavy chain (59 kDa) of MPO. (d) The effect of endoglycosidases on MPO binding. BAECs were pretreated with heparitinase, heparinase, and chondroitinase (all 8 mU·ml^–1) for 2 hours at 37°C, washed, and exposed to MPO (13 nM) at 4°C for 2 hours. Cell-associated MPO enzyme activity was then determined. *P < 0.05 for MPO alone versus pretreatment with heparin and enoxaparin (c) and versus heparitinase and heparinase pretreatment (d). (e) Binding competition analysis of MPO and xanthine oxidase (XO). BAECs were incubated with MPO (1 μg·ml^–1) and increasing concentrations of XO (0–100 μg·ml^–1, equivalent to 0–100 mU·ml^–1) for 2 hours at 37°C. Cells were harvested as above and MPO enzyme activity determined. Values represent mean ± SD.

Figure 3

Transcytosis of MPO by endothelial cells. (a) Intracellular localization and accumulation of MPO at the ECM. Cultured BAECs were grown to confluence and exposed to MPO (13 nM) for 2 hours. Cells were processed as described in Methods and incubated with mouse monoclonal anti–α-tubulin (red) and rabbit polyclonal anti-MPO (green). Laser confocal microscopy revealed intracellular MPO immunoreactivity within 2 minutes after MPO exposure and a high degree of MPO deposition at the subcellular matrix. Time = 0 minutes represents BAECs not exposed to MPO but stained with anti-MPO. ×63. (b) Colocalization of MPO and fibronectin (FN). BAECs treated as described for a were incubated with mouse monoclonal anti-fibronectin (red) and rabbit polyclonal anti-MPO (green). The nuclei (blue) were counterstained with DAPI. For assessing codistribution of MPO and fibronectin, images were merged (yellow). The control panel (Ctr) shows MPO-untreated cells stained with anti-MPO. The lower panel represents side-on views. ×63 (control, ×50). (c) Barrier function of endothelial cell monolayers exposed to MPO. BAECs were seeded on Transwell filters and exposed to increasing concentrations of MPO (13–130 nM) in the presence of TRITC-labeled dextran (4,400 Da) for 2 hours. MPO activity and TRITC fluorescence were assessed in the basolateral chamber. Treatment of the cells with H₂O₂ (100 μM) served as a control for increased permeability of TRITC-dextran. *P > 0.05 versus TRITC-labeled dextran alone. RFU, relative fluorescence units.

Figure 4

MPO-dependent tyrosine nitration of ECM fibronectin. (a) NO₂Tyr formation in cultured endothelial cell ECM proteins. Confluent BAEC monolayers were exposed to MPO (13 nM), and washed prior to NO₂^– (100 μM) and H₂O₂ (50 μM) addition. In some cases, cells were exposed to enoxaparin (Enox, 150 μg·ml^–1) and washed, followed by MPO exposure and no washing, before NO₂^– and H₂O₂ were added. Matrix-enriched protein fractions were isolated as described in Methods, separated by 4–20% SDS-PAGE gradient gels, and probed with mouse monoclonal anti-fibronectin and rabbit anti-NO₂Tyr. Protein staining with Coomassie blue confirmed that equal amounts of protein were electrophoretically resolved in each exposure condition. (b–d) NO₂Tyr formation in purified human fibronectin and its major fragments. Fibronectin (100 μg·ml^–1) (b) and the 30-, 45-, and 70-kDa fragments of fibronectin (d) were incubated with MPO (26 nM), μM NO₂^– (20–80 μM), and H₂O₂ (50 μM) in HBSS for 90 minutes. In some cases, fibronectin (c) and fibronectin fragments (d) were preincubated with enoxaparin (15 and 150 μg·ml^–1) for 45 minutes before MPO, NO₂^–, and H₂O₂ addition. Proteins were separated by SDS-PAGE electrophoresis (7.5% gels for fibronectin and 10% gels for fibronectin fragments) and then immunoblotted with rabbit polyclonal anti-NO₂Tyr.

Figure 5

Immunohistochemical distribution of NO₂Tyr in MPO- and ONOO^—exposed cultured endothelial cells. (a) MPO exposure. BAECs were exposed to MPO (13 nM) for 2 hours and washed, and NO₂^– (100 μM) and H₂O₂ (50 μM) were added for 90 minutes. (b) Peroxynitrite exposure. BAECs were exposed to an infusion of ONOO^– into culture medium for a final cumulative exposure of 100 μM. Cells were then fixed in 4% paraformaldehyde, permeabilized, and immunostained for fibronectin (red) and NO₂Tyr (green). Nuclei (blue) were counterstained with DAPI. For assessing codistribution of fibronectin and NO₂Tyr, images were merged (yellow). In control experiments, cells were incubated with MPO (13 nM) and H₂O₂ (50 μM) in the absence of added NO₂^–, and then immunostained for NO₂Tyr. Image acquisition was performed using laser confocal microscopy, with the lower panel depicting side-on views. ×63.

Figure 6

Immunolocalization of MPO, NO₂Tyr, and fibronectin in rat aortic rings. Rat aortic rings were exposed to MPO (65 nM) for 120 minutes, washed, and incubated with NO₂^– (100 μM) and H₂O₂ (50 μM) for 90 minutes. In some cases, vessel explants were preincubated with enoxaparin (150 μg·ml^–1) and washed prior to MPO exposure and omitting the washing step before H₂O₂ and NO₂^– addition. Tissue antigen distribution was visualized using rabbit polyclonal anti-MPO (green), mouse monoclonal anti-NO₂Tyr (red), and rabbit polyclonal anti-rat fibronectin (green). (a–c) Controls. Untreated vessel segments immunostained for MPO and NO₂Tyr (a), vessel segments incubated with MPO and stained for NO₂Tyr (b), and vessel segments treated with H₂O₂ and NO₂ stained for NO₂Tyr (c). (d–f) MPO and NO₂Tyr distribution. Immunoreactivity for MPO (d) and NO₂Tyr (e) colocalize, as shown when images were overlaid (f). (g–i) Effect of enoxaparin on NO₂Tyr formation. Untreated vessel stained for NO₂Tyr (g), MPO-catalyzed NO₂Tyr formation (h) was reduced when vessel explants were preincubated with enoxaparin (i). (j–l) Vascular FN and NO₂Tyr distribution. Subendothelial fibronectin immunoreactivity (j) colocalizes with NO₂Tyr immunoreactivity (k), as shown in the merged image (l). L, vessel lumen; M, media. Arrows indicate endothelial cells. ×100 (a–f, j–l); ×50 (g–i).

Figure 7

Nitrotyrosine/tyrosine ratio in liver of MPO^–/– mice after zymosan treatment. MPO^–/– and wild-type mice (MPO^+/+) received either intraperitoneal zymosan (n = 6), or sodium chloride (NaCl) as a control (n = 6). After 96 hours, the liver was perfused and processed for quantitative nitrotyrosine and tyrosine analysis as described in Methods. ^†P < 0.05, MPO^–/– zymosan-treated versus MPO+/+ injected with zymosan. *P < 0.05, MPO^+/+ zymosan-treated versus untreated MPO^+/+ control.