INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE NEUROPEPTIDE BRADYKININ is well known for its actions as an endothelium-dependent vasodilator. Bradykinin induces relaxation of vascular smooth muscle via stimulation of B₂ receptors, which in turn stimulates constitutively expressed endothelial nitric oxide (NO) synthase (eNOS) to produce NO, induces cyclooxygenase-dependent production of prostacyclin and other prostanoids, as well as superoxide, activates charybdotoxin-sensitive K⁺ channels, and induces the formation of epoxyeicosatrienoic acids by cytochrome P-450 epoxygenase (7, 8, 16, 18, 21, 65). In addition to its actions on arterial and arteriolar vascular smooth muscle, bradykinin also exerts powerful proinflammatory effects in postcapillary venules. For example, the neuropeptide causes the formation of interendothelial gaps and increased protein leakage; generates the release of endothelium-derived mediators from cultured endothelial cells that are chemotactic for neutrophils, eosinophils, monocytes, and pulmonary alveolar macrophages; induces the expression of endothelial adhesion molecules; and provokes leukocyte adherence to endothelial monolayers and postcapillary venules (1, 3, 11, 17, 25, 41, 43, 45, 53, 55-59).

In contrast to the large number of studies directed at elucidating the vasodilatory mechanisms of bradykinin in arteries and arterioles, comparatively less attention has been devoted to evaluation of the signaling pathways that underlie the proinflammatory effects of the neuropeptide in postcapillary venules. Indeed, our understanding of the mechanisms underlying the effect of bradykinin to produce leukosequestration is particularly inadequate. This latter question is especially intriguing in light of the fact that bradykinin induces the production of endothelium-derived mediators that exert powerful anti-inflammatory effects (e.g., NO, prostacyclin, and epoxyeicosatrienoic acids) (8, 16, 35, 37, 48, 51). Thus the aim of the present study was to explore the mechanisms whereby bradykinin disrupts microvascular barrier function and induces leukocyte rolling, adhesion, and emigration in postcapillary venules. To this end, we examined the effects of NOS or cyclooxygenase blockade, mast cell or cytoskeletal stabilization, oxidant scavenging, and xanthine oxidase, NADPH oxidase, cytochrome P-450 epoxygenase, or protein kinase C (PKC) inhibition on bradykinin-induced leukocyte adhesion and venular protein leakage. The molecular determinants of bradykinin-induced leukocyte/endothelial cell adhesive interactions were also interrogated by use of monoclonal antibodies directed against P-selectin and intercellular adhesion molecule-1 (ICAM-1).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical procedure. Male Sprague-Dawley rats (200-250 g) were maintained on a purified laboratory diet and fasted for 24 h before the experiment. The animals were initially anesthetized by intraperitoneal injection of pentobarbital sodium (65 mg/kg body wt). On attaining a surgical level of anesthesia, a tracheotomy was performed to facilitate breathing during the experiment. The right carotid artery was cannulated, and systemic arterial pressure was measured with a Statham P23A pressure transducer connected to the carotid artery catheter. Systemic blood pressure was continuously recorded with a personal computer (Macintosh 8500, Apple) equipped with an analog-to-digital converter (MP 100, Biopac Systems). The left jugular vein was also cannulated for administration of the anti-adhesion monoclonal antibodies. After these procedures, a midline abdominal incision was performed to allow for exteriorization and intravital microscopic examination of a section of the mesentery from the small intestine.

Intravital microscopy. The rats were positioned on a 20 × 30-cm Plexiglas board in a manner that allowed a selected section of mesentery to be placed over a glass slide covering a 4 × 3-cm hole centered in the Plexiglas. The mesentery was superfused at 2.5 ml/min with bicarbonate-buffered saline (BBS, pH 7.4) bubbled with a mixture of 5% CO₂-95% N₂ to reduce the oxygen tension to the physiological intraperitoneal level (40-50 mmHg). The exposed bowel wall was covered with BBS-soaked gauze to minimize tissue dehydration. The superfusate was maintained at 37 ± 0.5°C by pumping the solution through a heat exchanger warmed with a constant-temperature circulator (model 801, Fisher Scientific). Rectal temperature was monitored with an electrothermometer (4000A, YSI), and body temperature was kept at 36 ± 0.5°C with an infrared heat lamp. The Plexiglas board was mounted onto the stage of an inverted microscope (TMD-2S, Nikon), and a ×40 objective lens was used to observe the mesenteric microcirculation. The mesentery was transilluminated with a 12-V, 100-W direct current-stabilized light source. A video camera (VK-C150, Hitachi) connected to the microscope projected the image onto a color monitor (PVM-2030, Sony), and the images were recorded with a videocassette recorder (SLV-720HF, Sony). The time and date were displayed on both taped and live images with a date-time generator (WJ-810, Panasonic).

Experimental protocols. After a stabilization period of 30 min, images from the mesenteric preparation were recorded on videotape for 10 min (baseline recording). Thereafter, bradykinin was topically applied for 60 min to the exposed region of the mesentery via the superfusate. Video recordings were made during minute 20 to minute 30 and minute 50 to minute 60 of bradykinin superfusion. Hemodynamic measurements (i.e., venular diameter, erythrocyte velocity, and mean arterial blood pressure) and recording of fluorescence images (<10 s) were made at the end of each video recording. In the first series of experiments, the mesentery was superfused for 60 min with different concentrations of bradykinin (0.01, 0.1, or 1 µM). In an earlier study (59), we observed that higher concentrations of bradykinin (i.e., 10 µM), resulted in blood flow stasis in the venule under observation, an effect that was associated with a marked increase in the number of circulating platelet-leukocyte aggregates. Thus the maximal concentration of bradykinin we examined in the present study was 1 µM. In the second series of experiments, the mesentery was superfused with 1 µM bradykinin for 60 min in the presence of an NOS inhibitor, N^G-nitro-L-arginine methyl ester (L-NAME), at 10 µM (Sigma); a cyclooxygenase inhibitor, indomethacin, at 10 µM (Sigma); a B₂ receptor antagonist, HOE-140, at 1 µM (RBI); a platelet-activating factor (PAF) inhibitor, WEB-2086, at 10 µM (a gift from Boehringer Ingelheim); or a mast cell stabilizer, tranilast, at 10 µM (a gift from Kissei Pharmaceutical). Each drug was dissolved in BBS superfusate and was topically applied to the mesentery beginning 15 min before the baseline recording until the end of the experiment. The effect of phalloidin, an agent that stabilizes F-actin (25 µg/kg, Sigma, St. Louis, MO; administered 15 min before baseline measurements), was also tested. In a third group of experiments, the mesentery was superfused with 1 µM bradykinin for 60 min in the presence of superoxide dismutase [SOD, a highly specific superoxide anion scavenger that is largely confined to the extracellular space; 15,000 U/kg (Sigma) administered intravenously 15 min before baseline recordings in rats in which the renal vascular pedicles were ligated to prevent renal excretion], Mn(III)tetrakis(4-benzoic acid) porphyrin chloride [Mn-TBAP (Aldrich, Sheboygan, WI), a cell-permeant SOD mimetic, 10 µM via the superfusate], or N-(2-mercaptoproprionyl)-glycine [MPG (Sigma), a cell-permeant antioxidant, 300 µM via the superfusate]. In the fourth group of experiments, rats were treated with a xanthine oxidase [oxypurinol (Sigma), 20 µM via the superfusate], NADPH oxidase [PR-39 (a gift from Dr. C. R. Ross) dissolved in 0.5 ml saline and administered intravenously just after ligation of the renal vessels to produce a blood concentration of 5 µM, calculated assuming blood volume represents 6% of body weight], or a cytochrome P-450 2C9 epoxygenase [sulfaphenazole (Sigma), 10 µM, via the superfusate] inhibitor. Chelerythrine [1 µM (Sigma) via the superfusate, a pan-PKC inhibitor] or Go-6976 [10 nM (Calbiochem, La Jolla, CA) via the superfusate, a selective inhibitor of the calcium-dependent (classical) isotypes PKC- and PKC-BI] was used to assess the role of PKC and to define the isotype class involved in the proinflammatory effects of bradykinin. In a final group of studies, monoclonal antibodies directed against P-selectin (RMP-1, gift from Dr. D. N. Granger, 2.0 mg/kg) or ICAM-1 (1A-29, gift from Dr. D. N. Granger, 2.0 mg/kg) were administered as an intravenous bolus 15 min before the baseline recording in separate groups. Each group consisted of experiments conducted in 6-10 animals.

Statistical analysis. The data were analyzed with standard statistical analyses, i.e., ANOVA with Scheffé's (post hoc) test. All values are reported as means ± SE. Statistical significance was set at P < 0.05. The number of observations for each variable in each group was obtained from six to nine rats.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 shows the effects of superfusing the mesentery with different concentrations (0.01, 0.1, or 1 µM) of bradykinin on the number of adherent (Fig. 1A) and emigrated (Fig. 1B) leukocytes, number of rolling leukocytes (Fig. 1C), and venular protein leakage (Fig. 1D). While topically applied bradykinin at 0.01 µM did not induce leukocyte recruitment, higher concentrations of this neuropeptide (0.1 µM) significantly increased the number of adherent and emigrated leukocytes (Fig. 1, A and B). Although the number of rolling leukocytes also tended to increase at 0.1 µM bradykinin, the increase reached statistical significance only after 1 h of superfusion with the peptide at 1 µM (Fig. 1C). Topically applied bradykinin also significantly increased venular protein leakage at doses 0.1 µM (Fig. 1D).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Average number of adherent (A), emigrated (B), and rolling leukocytes (C) in single postcapillary venules of the rat mesentery after superfusing different concentrations (0.01, 0.1, and 1 µM) of bradykinin (BK). D: effects of increasing BK concentrations on venular protein leakage. Data show the values during baseline conditions and after 30 and 60 min of bradykinin superfusion. * Values that are statistically different from baseline at P < 0.05.

To investigate the role of NO, cyclooxygenase-derived prostanoids, bradykinin B₂ receptors, PAF, mast cell-derived mediators, microfilamentous actin disruption, P-selectin, and ICAM-1 in the effects of 1 µM bradykinin to induce leukocyte recruitment and venular protein leakage, we superfused the mesentery with L-NAME, indomethacin, HOE-140, WEB-2086, tranilast, or phalloidin or administered monoclonal antibodies directed against P-selectin (RMP-1) or ICAM-1 (1A29), respectively, as shown in Fig. 2. Bradykinin-induced leukocyte adhesion (Fig. 2A) and emigration (Fig. 2B) were completely prevented by coadministration of HOE-140, suggesting the involvement of B₂ receptor activation in these effects of bradykinin. Similar effects were noted with P-selectin and ICAM-1 immunoneutralization. In contrast, inhibition of NO synthesis with L-NAME or prostanoid production with indomethacin, PAF antagonism with WEB-2086, or mast cell stabilization with tranilast did not influence bradykinin-induced leukocyte adhesion and emigration. Stabilization of microfilamentous actin with phalloidin prevented leukocyte emigration without modifying the increase in leukocyte adhesion induced by bradykinin. The effect of bradykinin to increase the number of rolling leukocytes was also compared in the absence (control) and the presence of various agents described above (Fig. 2C). The bradykinin-induced increase in leukocyte rolling was attenuated by HOE-140, WEB-2086, phalloidin, or tranilast and was completely prevented by P-selectin immunoneutralization. In contrast, L-NAME, indomethacin, and anti-ICAM-1 monoclonal antibodies did not prevent the increase in the number of rolling leukocytes caused by bradykinin. With the exception of HOE-140 or phalloidin, both of which prevented bradykinin-induced venular protein leakage, none of the other treatments described above influenced the effect of the peptide to disrupt the microvascular barrier (Fig. 2D).

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Effect of superfusing the mesentery with BK (1 µM) in the absence (BK alone) or presence of NG-nitro-L-arginine methyl ester [L-NAME; +nitric oxide synthase (NOS) Block], indomethacin [+cyclooxygenase (COX) Block], HOE-140 (+B2 Antagonist), WEB-2086 [+platelet-activating factor (PAF) Antagonist], or tranilast (+Mast Cell Stab) on the no. of adherent (A), emigrated (B), and rolling (C) leukocytes and venular protein leakage (D). In additional experiments, phalloidin (+Actin Stab), or monoclonal antibodies directed against functional epitopes on P-selectin (+Anti-P-sel) or intercellular adhesion molecule (ICAM)-1 (+Anti-ICAM-1) were administered intravenously as described under METHODS. Data show the values during baseline conditions and after 30 and 60 min of BK superfusion alone or in combination with the agents listed above. * Values statistically different from baseline at P < 0.05.

The data depicted in Fig. 3 indicate that bradykinin-induced oxidant generation contributes to its proinflammatory effects. Although SOD did not alter leukocyte rolling, adhesion, or emigration, superfusing the mesentery with the cell-permeant antioxidants Mn-TBAP and MPG attenuated these responses to bradykinin (Fig. 3, A-C). A similar pattern was noted with regard to the effect of these inhibitors on bradykinin-induced venular protein leakage (Fig. 3D). To determine if xanthine oxidase or NADPH oxidase represented important sources of the oxidant-induced proinflammatory effects of bradykinin, these enzymes were inhibited with oxypurinol or PR-39, respectively. Neither agent was effective in attenuating the effects of bradykinin to increase leukocyte rolling, adhesion, and emigration or disrupt the microvascular barrier (Fig. 3, A-D). However, blockade of cytochrome P-450 epoxygenase activity (+Sulfaphenazole), which produces functionally significant concentrations of reactive oxygen species in response to ligation of bradykinin receptors (19), prevented the inflammatory responses to bradykinin (Fig. 3, A-D).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Effect of BK (1 µM) in the absence (BK alone) or presence (+) of superoxide dismutase (SOD), N-(2-mercaptoproprionyl)-glycine (MPG), Mn(III)tetrakis(4-benzoic acid) porphyrin chloride (TBAP), oxypurinol, PR-39, or sulfaphenazole on the no. of adherent (A), emigrated (B), and rolling (C) leukocytes and venular protein leakage (D). Data show the values during baseline conditions and after 30 and 60 min of BK superfusion. * Values statistically different from baseline (P < 0.05). # Values statistically different from corresponding value in BK-alone group (P < 0.05).

Because oxidants activate PKC and this family of kinases has been implicated in the effect of bradykinin to increase microvascular permeability (6, 22, 46, 54), we sought to determine whether this kinase family also participated in bradykinin-induced leukocyte infiltration. Superfusion of the mesentery with chelerythrine, a pan-PKC antagonist, prevented the effects of bradykinin to increase leukocyte/endothelial cell adhesive interactions and venular protein leakage (Fig. 4, A-D). Whereas the latter observations support a role for PKC in the signaling cascade induced by bradykinin, they do not provide insight regarding the class of PKC isoforms involved because chelerythrine blocks the activity of all PKC isotypes. To determine the role of calcium-dependent (classical or conventional) PKC isotypes in the proinflammatory effects of bradykinin, we treated the mesentery with Go-6976, a highly selective antagonist of the classical isoforms PKC- and PKC-BI that exhibits no activity toward the novel isotypes PKC- or PKC-, even at millimolar concentrations (42). Go-6976 was as effective as chelerythrine in preventing the effects bradykinin to increase venular protein leakage and leukocyte rolling, adhesion, and emigration (Fig. 4, A-D).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effect of BK (1 µM) in the absence (BK alone) or presence of the protein kinase C antagonists chelerythrine or Go-6976 on the number of adherent (A), emigrated (B), and rolling (C) leukocytes and venular protein leakage (D). Data show the values during baseline conditions and after 30 and 60 min of BK superfusion. * Values statistically different from baseline (P < 0.05). # Values statistically different from corresponding value in BK alone group (P < 0.05).

The effect of bradykinin superfusion at 0.01, 0.1, or 1 µM on venular diameter, erythrocyte velocity, and mean arterial blood pressure is summarized in Table 1. Topical application of bradykinin did not alter the venular diameter or mean arterial blood pressure at any dose but significantly decreased erythrocyte velocity when superfused at 1 µM (Table 1). Venular diameter and mean arterial blood pressure were not modified by bradykinin alone or bradykinin in combination with any of the treatments, averaging approximately 30 µm and 140 mmHg, respectively, in each group at all time points during the protocols (data not shown). Erythrocyte velocity averaged ~3.25 mm/s in all groups under basal conditions and gradually decreased by ~33% in every group by the end of the protocol (i.e., after 60 min of bradykinin superfusion) except in those animals treated with HOE-140 and tranilast, where erythrocyte velocity did not change over the course of the experiments, averaging 3.14 ± 0.18 and 2.64 ± 0.23 mm/s, respectively, after 60 min of bradykinin superfusion. These results suggest that changes in venular wall shear rate (which is proportional to venular erythrocyte velocity times venular diameter) do not account for the effects of the pharmacological agents on bradykinin-induced leukocyte adhesion (10, 52).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bradykinin is a vasoactive proinflammatory neuropeptide that induces relaxation of vascular smooth muscle in arteries and arterioles and promotes adhesion molecule expression, leukocyte sequestration, and the formation of interendothelial gaps and protein extravasation in postcapillary venules (1, 3, 6-8, 11, 16-18, 21, 25, 41, 43, 45, 46, 53-55, 58, 59, 65). While the mechanisms underlying the vasorelaxant properties of this peptide have been extensively studied, the cell signaling pathways that mediate the powerful proinflammatory effects of bradykinin in postcapillary venules have received less attention. Indeed, the effect of the neuropeptide to induce leukocyte infiltration has only recently been recognized (45, 58, 59), and little information is available regarding the mechanisms underlying this response. Even though the signaling pathways that mediate the increased microvascular permeability induced by bradykinin are better understood (1, 3, 6, 11, 17, 41, 43, 46, 54), there are still large gaps in our understanding of the mechanism involved in this inflammatory response to the peptide. Thus the major aims of this study were to begin to identify the signaling pathways whereby bradykinin induces leukocyte rolling, adhesion, and emigration and to expand our understanding of the mechanisms underlying the effect of the peptide to increase protein extravasation.

We observed that superfusion of the mesentery with bradykinin for 60 min increased the number of rolling, firmly adherent (stationary), and emigrated leukocytes in a manner that appeared to be concentration and time dependent (Fig. 1). The bradykinin-induced leukocyte/endothelial cell adhesive interactions were associated with a decrease in venular erythrocyte velocity without a change in mean arterial blood pressure or venular diameter. These results are qualitatively similar to those we reported in an earlier study (59) in which bradykinin was applied to the mesentery for a much shorter time period (15 vs. 60 min in the present study). In this earlier study, we demonstrated that leukocyte adhesion induced by bradykinin was mediated by a mechanism that was dependent on B₂ receptor activation and the formation of PAF or PAF-like lipids (59). These observations were consistent with the results of in vitro studies indicating that bradykinin releases PAF or PAF-like lipids from endothelial monolayers and macrophages and induces leukocyte adhesion to cultured endothelium (44, 57, 61). As observed with short-term (15 min) exposure to bradykinin in our earlier report (59), bradykinin B₂ receptor activation appears to be essential for the increased leukocyte rolling, adhesion, and emigration and the venular protein leakage noted with more prolonged exposure to bradykinin (60 min in the present study) because HOE-140 treatment completely prevented these responses (Fig. 2). As we had also shown earlier with 15 min of bradykinin exposure, the PAF antagonist WEB-2086 attenuated leukocyte rolling induced by 60-min superfusion with the peptide. However, in stark contrast to our earlier results with short-term bradykinin exposure (59), WEB-2086 was not effective in preventing the adherence and emigration of leukocytes induced by more prolonged exposure to this neuropeptide (Fig. 2). The divergent results reported in the present vs. our previous study (59) suggest that bradykinin-induced leukocyte adhesion and emigration occurs via a time-dependent mechanism with the initial phase (15 min) being PAF dependent and later phases (30-60 min) being PAF independent.

We examined the role of mast cell-derived mediators by superfusing mesenteries with tranilast, a mast cell stabilizer (47). This notion was based on the facts that bradykinin can cause mast cell degranulation and many of the secreted substances (e.g., histamine and serotonin) are not only vasoactive but also induce leukocyte adherence and increase microvascular permeability (39, 40, 62). Tranilast attenuated the increase in the number of rolling leukocytes, a result that indicates that the agent remained effective over the 60-min protocol (Fig. 2). However, mast cell stabilization failed to prevent leukocyte adhesion and emigration induced by bradykinin (Fig. 2). While these results indicate that the degranulation of mast cells may lead to the release of mediators that promote leukocyte rolling in postcapillary venules of the rat mesentery, mast cell-derived mediators do not appear to play an important role in the bradykinin-induced leukocyte adhesion and emigration.

Bradykinin has been reported to stimulate the formation of superoxide by cultured endothelial cells (27, 60), and inhibitor studies indicate that superoxide generation contributes to the vascular effects of bradykinin in vivo (64). Based on these observations, we hypothesized that bradykinin-induced superoxide formation may play a role in its proinflammatory effects in our model. This postulate is especially appealing in light of the facts that superoxide has been implicated in the increased leukocyte adhesion induced by ischemia/reperfusion (23, 24), and xanthine oxidase-derived superoxide (and other secondarily-derived oxidants) promote the adherence of granulocytes to the microvascular endothelium (63). To address this issue, rats were treated with either SOD (a highly specific superoxide scavenger), a cell-permeant SOD mimetic (Mn-TBAP), or MPG, an antioxidant that also readily crosses cell membranes. While SOD was not effective in preventing the proadhesive effects of bradykinin, Mn-TBAP and MPG markedly attenuated the increase in leukocyte adhesion and emigration induced by the peptide (Fig. 3). These apparently disparate results may be explained by the fact that SOD does not readily cross cell membranes (and thus is largely confined to the extracellular space after intravenous administration) whereas both Mn-TBAP and MPG can gain access to the intracellular compartment. Thus our results are consistent with the view that bradykinin induces leukocyte adhesion and emigration by a mechanism that involves oxidant generation at an intracellular site. This conclusion is supported by the work of Wolin and co-workers (67), who demonstrated that the vasodilator responses to bradykinin do not involve extracellular oxidant generation. Moreover, bradykinin-induced vascular relaxation is reduced in vessels exposed to an inhibitor of cytosolic Cu/Zn-SOD and enhanced by cell-permeant SOD mimetics (64), observations that also point to an intracellular oxidant source being activated by bradykinin.

In light of the aforementioned findings supporting a role for oxidants in the proadhesive effects of bradykinin, we next attempted to determine the source for their production. It is well known that bradykinin can promote the formation of superoxide as a result of cyclooxygenase activation (66). In addition, NOS is activated by ligation of bradykinin B₂ receptors and can produce superoxide under certain conditions (66). Moreover, it is also possible that the generation of chemotactic stimuli may occur secondary to concomitant activation of NOS and cyclooxygenase by bradykinin. That is, bradykinin may induce the simultaneous formation of NO and superoxide by stimulating these enzymes. NO and superoxide may interact to form the powerful oxidants, peroxynitrite and hydroxyl radicals, which may exert proinflammatory effects in their own right but may also participate in the formation of chemotactic stimuli. Indeed, bradykinin has been shown to induce the simultaneous formation of superoxide and NO in canine basilar arteries (64). However, treatment with indomethacin (cyclooxygenase inhibitor) or L-NAME (NOS inhibitor) proved to be ineffective in inhibiting the bradykinin-induced leukocyte recruitment (Fig. 2).

Although our L-NAME data argue against a role for NO in bradykinin-induced venular protein leakage in the rat mesentery, work conducted in the hamster cheek pouch indicates that this NOS antagonist prevents the effect of the neuropeptide to cause protein extravasation (17, 43). While it is not clear why our results in rat mesentery differ from those reported in hamster cheek pouch, it is likely that differences in tissues, species, time-course for bradykinin exposure, and/or method for assessing endothelial barrier function may account for the discrepant results.

Because our results did not support a role for the generation of radical species from cyclooxygenase or NOS, we examined the contribution of two other well-known sources of superoxide, xanthine oxidase and NAD(P)H oxidase, to the proinflammatory effects of bradykinin. Both enzymes are present in the vascular wall, and NADPH oxidase is activated in stimulated neutrophils (26, 50). In addition, there is some evidence for xanthine oxidase activation by bradykinin (15). However, this latter notion is controversial, with other studies indicating that the peptide does not promote an increase in xanthine oxidase activity (20). Our results fail to support a role for either enzyme because neither xanthine oxidase inhibition with oxypurinol nor treatment with PR-39, a proline/arginine-rich peptide that inhibits NADPH oxidase, was effective in preventing the proinflammatory effects of bradykinin (Fig. 3). Thus xanthine oxidase and NAD(P)H oxidase do not appear to play a role in generating the oxidant species that mediate the proinflammatory effects of bradykinin.

Recent work indicates that bradykinin elicits the production of reactive oxygen species in endothelial cells via activation of cytochrome P-450 2C9 (CYP 2C9) (19). To test for a role for this endothelium-derived hyperpolarizing factor synthase in the inflammatory responses to bradykinin, we evaluated the effect of a selective CYP 2C9 inhibitor, sulfaphenazole, on bradykinin-induced leukocyte/endothelial cell adhesive interactions and venular protein leakage. Coincident treatment with sulfaphenazole prevented the inflammatory responses to bradykinin (Fig. 3). While these observations support the notion that CYP 2C9 may be a source of the reactive oxygen species that initiate the proinflammatory effects of bradykinin, this enzyme also catalyzes the formation of epoxyeicosatrienoic acids (19, 48). However, it is unlikely that epoxyeicosatrienoic acids contribute to the proinflammatory effects of bradykinin because these CYP epoxygenase-derived eicosanoids have been shown to reduce cytokine-induced adhesion molecule expression and prevent leukocyte adhesion to the vascular wall (48). These observations, when coupled with our data, suggest that the proinflammatory effects of bradykinin depend on CYP epoxygenase-induced oxidant generation, an effect that may outweigh the anti-inflammatory actions of epoxyeicosatrienoic acids that are formed coincidentally as a result of the catalytic activity of this enzyme.

Demonstrating a role for reactive oxygen species in the proinflammatory effects of bradykinin suggested the possibility that PKC-dependent phosphorylation may participate in the signaling cascade induced by B₂-receptor activation because oxidants can activate this family of enzymes (22). Strong evidence has accumulated indicating that PKC plays an important role in mediating the effect of bradykinin to increase microvascular permeability. For example, PKC inhibition with sphingosine or H-7 markedly attenuated the bradykinin-induced appearance of leaky sites and increased macromolecule (dextran 70) clearance in the cheek pouch microcirculation (46). Similar results have been reported in cultured endothelial cells exposed to bradykinin in that depletion of PKC by long-term phorbol ester exposure or inhibition of the enzyme with staurosporine attenuated the increase in albumin permeation induced by the peptide (6). Our results demonstrating that administration of the PKC inhibitors chelerythrine or Go-6976 abrogates the increase in venular protein leakage induced by bradykinin confirms and extends these earlier observations (Fig. 4). An important new finding of the present study is that bradykinin-induced PKC activation also participates in the effect of the peptide to recruit leukocytes. Our studies also provide novel insight regarding the PKC isotypes involved in the proinflammatory effects of bradykinin. That is, our data indicate that the effects of bradykinin involve the calcium-dependent or classical isotypes of the enzyme. The latter notion is based on the observation that Go-6976, a PKC antagonist that exhibits a high degree of specificity for the classical isotypes PKC- and PKC-BI but demonstrates no inhibitory activity toward the novel, calcium-independent isoforms (e.g., PKC- or PKC-) even at millimolar concentrations (42), was as effective as chelerythrine in preventing the effects of bradykinin to promote leukocyte adhesion and microvascular barrier dysfunction (Fig. 4). This conclusion is supported by the observation that bradykinin causes the translocation of PKC- from the cytosolic to membrane fraction in cultured human umbilical vein endothelial cells (54). The demonstrated role for PKC is also consistent with our tranilast data in that bradykinin-induced mast cell mediator release occurs by a PKC-independent mechanism (12). Thus if mast cells were involved in the effects of bradykinin to increase venular protein leakage and leukocyte adhesion, which our data argue against, then one would have expected that PKC inhibition would be ineffective in preventing these actions of the peptide, which was not the case.

Although the molecular targets of bradykinin-induced PKC--dependent phosphorylation are uncertain, it has been suggested that cytoskeletal proteins may represent downstream elements in the activation process (54). This notion is appealing with regard to work indicating that cytoskeletal proteins play an important role in the activation and expression of adhesion molecules and in the regulation of endothelial barrier function (3, 5, 30, 41). To address this issue, we administered phalloidin, a bicyclic heptapeptide derived from the green-capped mushroom Amanita phalloides, which acts to induce the assembly of F-actin as well as to stabilize existing actin filaments in microfilamentous cytoskeletal elements. We have previously exploited these properties of phalloidin to evaluate the role of endothelial cytoskeletal reorganization in the leukocyte infiltration and microvascular barrier disruption induced by PAF, leukotriene B₄, N-formyl-methionyl-leucyl-phenylalanine, or ischemia/reperfusion (3, 5, 30). A significant new finding of the present study is that phalloidin attenuated leukocyte rolling and completely prevented leukocyte emigration and albumin leakage induced by bradykinin, without affecting the firm (stationary) adhesive interactions between white cells and the postcapillary venular endothelium that occur as a result of B₂ receptor activation (Fig. 2).

We can only speculate as to the mechanism whereby phalloidin attenuated bradykinin-induced leukocyte rolling and prevented leukocyte emigration and microvascular barrier disruption. Because leukocyte rolling is dependent on P-selectin expression, stabilizing F-actin in endothelial cells may limit bradykinin-induced mobilization of this adhesive ligand from Weibel-Palade bodies to the cell surface (5). The fact that phalloidin prevents the changes in endothelial cell shape required for diapedesis may explain its ability to prevent bradykinin-induced leukocyte emigration and microvascular barrier disruption (3, 5, 30). Stabilizing F-actin in endothelial cells may also prevent alterations in the organization of junctional proteins, thereby limiting diapedesis and venular protein leakage (3). Because leukocyte rolling has been suggested to be mandatory for subsequent adhesion and emigration, it might be viewed as surprising that phalloidin failed to influence bradykinin-induced stationary adhesion, given its effect to reduce rolling by 60-70%. However, Kubes (33) has demonstrated that leukocyte rolling must be inhibited by >90% to achieve significant reductions in leukocyte adhesion.

Our results also indicate that bradykinin-induced leukocyte adhesion and emigration depend on the expression of specific adhesion molecules, P-selectin and ICAM-1. Although P-selectin antibody treatment blocked bradykinin-induced leukocyte rolling as well as leukocyte adhesion and emigration, immunoneutralization of ICAM-1 blocked only stationary adhesion and emigration (Fig. 2). Because leukocyte rolling is a prerequisite to establishing stationary adhesive interactions, these observations are compatible with the notion that bradykinin-induced leukocyte rolling is P-selectin dependent while firm adhesion is mediated by ICAM-1. Consistent with this concept is the observation that intraperitoneal administration of bradykinin upregulates P-selectin expression in rat gastrointestinal organs (55). Bradykinin-induced oxidant formation may contribute to the expression of P-selectin because surface expression of this adhesive glycoprotein can be upregulated by reactive oxygen species (2).

Because there is persuasive evidence suggesting that there is a causal link between mediator-induced leukocyte emigration and vascular protein leakage (38), it was surprising that immunoneutralization of P-selectin or ICAM-1, an approach that completely abolished the leukocyte adhesion and emigration induced by bradykinin, did not inhibit venular albumin leakage (Fig. 2). These findings indicate that the microvascular barrier disruption caused by bradykinin occurs by a mechanism that is leukocyte independent. This concept is supported by in vitro studies wherein bradykinin induces a reorganization of cytoskeletal elements and disrupts barrier function of endothelial monolayers in the absence of neutrophils (3, 54). In addition, Norman et al. (49) have shown that monoclonal antibodies directed at CD18 or ICAM-1 do not prevent edema formation produced by bradykinin in rabbit skin.

Several methodological considerations warrant some attention. For example, it may be argued that L-NAME, WEB-2086, tranilast, PR-39, and/or SOD may have become ineffective over the prolonged protocol used in the present study. However, this potential explanation for the lack of effect of these agents on leukocyte adhesion is unlikely for at least two reasons. First, L-NAME, WEB-2086, and tranilast were continuously applied to the mesenteric venule under observation throughout the protocol. In addition, the circulating half-life of SOD is increased from 8 min to nearly 2 h by ligating the renal vascular pedicle (32) before its administration, as we did. Renal vascular ligation would also be expected to increase the circulating half-life of PR-39 (31). Thus it is not likely that drug metabolism or renal elimination accounts for the lack of effect of these agents. Second, using the same model (rat mesentery) and protocol for antagonist delivery (continuous application via the superfusate) that we have employed in the present study, Granger and co-workers (4, 34, 35) have been able to show that superfusion with WEB-2086 prevents leukocyte adhesion and venular protein leakage induced by 60-min (the same time frame over which we observed the responses to bradykinin) NOS inhibition or by ischemia followed by 60-min reperfusion. Thus WEB-2086 effectively prevented PAF-induced effects associated with ischemia/reperfusion and NOS inhibition in the rat mesentery over the same period of time that the rat mesentery was exposed to bradykinin in our studies, where the inhibitor had no effect. Similar arguments apply to the data obtained with L-NAME, SOD, oxypurinol, PR-39, indomethacin, and mast cell stabilization (28, 29, 31, 34, 50).

In summary, exposing the rat mesentery to bradykinin at concentrations 0.1 µM for 60 min increased the number of rolling, firmly adherent, and emigrated leukocytes in single postcapillary venules. These bradykinin-induced leukocyte/endothelial adhesive interactions occurred by a B₂-receptor-initiated, oxidant- and PKC-mediated, P-selectin- and ICAM-1-dependent mechanism. Similar concentrations of bradykinin also produced venular protein extravasation. While this latter effect was also initiated by B₂ receptor activation, involved oxidant generation, PKC activation, and disruption of microfilamentous actin, bradykinin-induced microvascular barrier disruption was independent of leukocyte adherence and emigration.