Protease-Activated Receptor-2 Activation Induces Acute Lung Inflammation by Neuropeptide-Dependent Mechanisms ¹

Protease-activated receptors (PARs) ³ serve to link tissue damage to a variety of cellular responses (1, 2, 3, 4). Roles for PARs in hemostasis and thrombosis are established, and roles in inflammation are emerging (5, 6, 7, 8). Mammals express four PARs, of which three, PAR1, PAR2, and PAR4, are known to play important signaling roles. These three receptors are all expressed on endothelial cells, and superfusion of blood vessels with thrombin or PAR-activating peptides (PAR-AP) induces leukocyte rolling, adhesion, and extravasation (9, 10, 11). These and other responses to PAR activation are prominent features of lung injury, and PARs are widely expressed in lung. However, the role of PARs in this tissue has not been completely explored. PAR1 and PAR2 are expressed on bronchial epithelial (12) and smooth muscle cells (13, 14, 15), as well as vascular endothelial and smooth muscle cells (16). A type II alveolar epithelial cell line (A549) also expresses PAR2 (17). PAR2 is present on capsaicin-sensitive sensory afferent nerves (18), which are well distributed in the lung (19, 20). Similarly, PAR2 is also expressed by a subset of sensory neurons in the spinal dorsal root ganglia (DRG) (5, 21), and the lung is innervated by their afferent fibers. In the skin, PAR2 activation has been reported to trigger neurogenic inflammation characterized by plasma extravasation and neutrophil infiltration (5).

To probe the potential roles of PARs in lung, we instilled PAR-APs into the airspaces of the lungs in anesthetized, ventilated, and spontaneously breathing mice. PAR2-dependent effects of PAR2-AP on airway pressure, lung vascular, and epithelial permeability to protein, extravascular lung water, arterial blood gases, and leukocyte infiltration were measured. Pharmacologic ablation of sensory neurons and neuropeptide antagonists was used to determine the mechanisms of PAR2-mediated lung pathophysiology.

Mice (C57BL/6J; 8 wk old) were purchased from The Jackson Laboratory. Generation of mice deficient in PAR2 has been described elsewhere (11). All procedures were approved by the Animal Committee of University of California.

The PAR-APs, including TFLLRN-NH₂ (PAR1-AP), SLIGRL-NH₂ (PAR2-AP), AYPGKF-NH₂ (PAR4-AP), and control SLIGRL peptide (LSIGRL) were synthesized as carboxyl amides and purified by reverse-phase HPLC (AnaSpec). The PAR-APs and LSIGRL were dissolved in HEPES-buffered Hanks’ at concentration of 50 mM. Capsaicin, myeloperoxidase (MPO), malondialdehyde (MDA), CGRP8–37 (calcitonin gene-related peptide receptor antagonist), and other reagents were purchased from Sigma-Aldrich. Spantide III (a NK₁ receptor antagonist, sequence: H-d-Lys(nicotinoyl)-Pro-b-(3-pyridyl)-Ala-Pro-3,4-dichloro-d-Phe-Asn-d-Trp-Phe-b-(3-pyridyl)-d-Ala-Leu-Nle-NH₂) and GR87389 (a NK₂ receptor antagonist, sequence: Boc-Ala-Ala-d-Trp-Phe-d-Pro-Pro-Nle-NH₂) were purchased from Bachem. The neuropeptide antagonists were dissolved in 0.9% saline.

To deliver PAR-APs during ventilation, a lateral vent was made in a tracheal cannula at the proximal end of the carina and connected with a PE 10 by superglue. An insulin syringe containing 25 μl of PAR-APs (50 mM), 25 μl of LSIGRL (50 mM), or Hanks’ was connected with PE-10 and was injected in 1 min, followed by 25 μl of air. To prevent reflux of the instillate, mice were placed in a supine position on a board with an angle of 45°. In nonventilated mice, anesthesia was induced with ketamine (90 mg/kg) and xylazine (10 mg/kg). The mouse was suspended with incisors attached to a ∼60° wood support by 3/0 suture. A cold-light source (Dolan-Jenner Industries) with two 25-inch flexible fiber-optic arms allowed transillumination to visualize glottis and vocal cords to deliver 25 μl of PAR-APs (50 mM), 25 μl of LSIGRL (50 mM), or Hanks’ into the airspaces (22). The investigator was blinded to the genotype of the mice and the instillate.

Mice were anesthetized by ketamine (90 mg/kg) and xylazine (10 mg/kg) given i.p. A tracheotomy was done with implantation of a tracheal cannula, a PE-90, with a lateral vent connected with PE 10 to deliver the PAR-APs. Pancronium (0.5 mg/kg) was injected i.p. to prevent spontaneous breath during mechanical ventilation. Mice were ventilated with 30% O₂ at 130 breaths/min and 7 ml/kg tidal volume, using a rodent respirator (MINI vent, type 845; March-Hugstetten). Body temperature was maintained at 37°C with a heating lamp and monitored by a rectal thermometer (Fisher Scientific). The right jugular vein was catheterized for administration of radioactive label (0.05 μCi of ¹²⁵I-labeled albumin (¹²⁵I-albumin; Merck Frosst Labs) or vehicle. Airway pressure was recorded by Biopac system (BIOPAC Systems) through a lateral passage connected with the tracheal cannula (PE-90). The airway peak pressure and plateau pressure at 30 min (after subtracting the baseline) were used to compare the effects of the different PAR-APs. Blood samples were withdrawn from the carotid arterial catheter at the end of the experiment and measured by a blood gas analyzer (Diamond Diagnostics).

The lungs were removed, counted in the gamma counter (Packard Instrument), weighed, and homogenized (after addition of 1 ml of distilled water). The blood was collected through right ventricle puncture. The homogenate was weighed, and a portion was centrifuged (12,000 rpm, 8 min) for assay of hemoglobin concentration in the supernatant. Another portion of homogenate, supernatant, and blood was weighed and then desiccated in an oven (60°C for 24 h) for gravimetric determination of extravascular lung water. The lung wet-to-dry weight ratio (lung W/D ratio) was calculated by standard formula (23, 24). ELW was calculated by: (lung W/D ratio_experimental× lung dry weight_experimental − lung W/D ratio_normal× lung dry weight_normal) × 1000 (μl). Lung EPE (index of lung vascular permeability to protein) were calculated as the counts of ¹²⁵I-albumin in the lung tissue divided by the counts of ¹²⁵I-albumin in the plasma (23, 24).

The activity of myeloperoxidase and the levels of malondialdehyde were measured by standard methods (25, 26).

We used the methods as previously described (27). The white blood cells were counted by a counter (Beckman Coulter). Blood cell smear was made using cytospin (Thermo Shandon). The slides were visualized using Wright-Giemsa staining (Fisher Scientific). Neutrophils and macrophages were identified by a certified laboratory technologist in a blinded fashion. Protein concentration in the BAL was determined by a Bio-Rad protein assay (Bio-Rad). MIP-2 (analog of IL-8) in the BAL was measured by ELISA (R&D Systems). For measurement of substance P in the BAL, mice were euthanized at 30 min and the lungs were lavaged with 0.4 ml of chilled PBS (containing 500 kallikrein inhibitor U/ml aprotonin). Substance P levels in the BAL were assessed by ELISA (R&D Systems).

After euthanasia by inhalation of halothane, the chest and abdomen were rapidly opened and the base of the heart was clamped to prevent escape of the pulmonary blood volume. The thoracic organs were removed en bloc, and 10% Formalin was instilled through the trachea at a pressure of 25 cm H₂O. After 72-h fixation, lungs were embedded in paraffin, and 5-μm sections were cut and stained with H&E. Using light microscopy (×400), a counting grid (1 mm² of the magnified field) was used as a field for neutrophil counting. A sequestered neutrophil was defined as a neutrophil within a pulmonary capillary. An emigrated neutrophil was defined as a neutrophil within the extravascular (interstitial and alveolar) compartments. On each slide, we randomly selected three to four counting fields and calculated the mean value (28, 29). The investigator who reviewed the pathology was blinded to the experimental groups.

For ablation of the C-fiber sensory neurons, under halothane anesthesia, individual mice received three consecutive doses of capsaicin (25, 50, and 50 mg/kg, s.c.) dissolved in vehicle (10% ethanol, 10% Tween 80, sterile physiological saline) over 32 h (at 0, 6, and 32 h, respectively; total dose 125 mg/kg) or an equivalent volume of vehicle. Mice were studied 10 days after the last injection (5).

Spantide III (NK1 receptor antagonist; 2 mg/kg) (30), GR87389 (NK2 receptor antagonist; 2 mg/kg) (31), or CGRP8–37 (CGRPα1 receptor antagonist; 10 μg/kg) (5) was given i.p. 25∼30 min before PAR2 stimulation to block the effects of neuropeptides. The respective control received the same amount of vehicle (0.9% saline).

Statistics were done by SPSS software. An unpaired t test was used unless there were multiple comparisons, in which case we used ANOVA (significance level set at p ≤ 0.05). The results are presented as mean ± SD.

PAR2-AP instilled into the airspaces increased peak and plateau airway pressure compared with Hanks’ control and control peptide (Fig. 1). PAR2-AP also increased ELW and lung vascular permeability to protein at 30 min in a receptor-dependent manner (Fig. 2, A and B). PAR2-AP also induced respiratory acidosis, hypoxemia, and hypercapnia (Table I). The effect of PAR2 activation on airway pressure and ELW was dose dependent (Fig. 2, C and D). Airspace PAR1-AP and PAR4-AP did not significantly affect any of these parameters.

Plasma MPO activity (an index of neutrophil degranulation) and MDA concentration (an indicator of lipid peroxidation) were significantly increased in wild-type mice that received the PAR2-AP compared with wild-type mice instilled with Hanks’, control peptide, or PAR2 knockout mice instilled with PAR2-AP (Fig. 2, E and F). Airspace PAR1- and PAR4-APs did not affect these parameters.

To study later effects of PAR2-AP on lung inflammation, mice were instilled with Hanks’, LSIGRL, or PAR2-AP, and the protein content, leukocyte numbers, and MIP-2 levels in the BAL were examined at 4 h. The protein concentration and leukocyte numbers in the BAL were increased compared with controls (Fig. 3, A and B). Neutrophils were increased 12-fold in the PAR2-AP-instilled group compared with control peptide (Fig. 3,C) and macrophages were slightly increased (Fig. 3,D). PAR2 activation caused an 8-fold increase of MIP-2 levels (Fig. 3,E). To study neuropeptide release after airspace exposure of PAR2-AP, the lungs were lavaged at 30 min. PAR2 activation caused a 5-fold increase in substance P levels in the BAL compared with the control peptide (Fig. 3 F).

To examine histological changes induced by PAR2-AP, mice were instilled with Hanks’, LSIGRL, or PAR2-AP and euthanized at 30 min and 4 h. At 30 min, the alveolar septa were wider (more edema) and more neutrophils were sequestered in the pulmonary microcirculation in the PAR2-AP group (Fig. 3, G and I). At 4 h, more emigrated neutrophils were present in the PAR2-AP group (Fig. 3, H and J).

To determine whether airway responses to PAR2 activation were mediated by neuropeptides, we pretreated mice with capsaicin to ablate C-fibers before instillation of Hanks’ or PAR2-AP. Mice used to study immediate (30 min) and later (4 h) responses were ventilated. The elevated peak pressure and plateau airway pressures at 30 min induced by PAR2 activation in sensory neuron-ablated mice were reduced compared with intact control mice (Fig. 4, A and B). The increase in ELW and EPE in the capsaicin group was also decreased compared with control group at both 30 min (Fig. 4, C and D) and 4 h (Fig. 4, E and F) after PAR2-AP instillation.

To determine which neuropeptides contribute to airway constriction and lung injury, we pretreated mice with neuropeptide receptor antagonists or corresponding vehicle before instillation of PAR2-AP or Hanks’. NK2 receptor blockade (GR87389) reduced the increases in peak and plateau airway pressure compared with the vehicle, whereas NK1 (Spantide III) and CGRP receptor blockade did not affect airway pressures (Fig. 5, A and B). Both NK1 (Spantide III) and NK2 receptor blockade (GR87389) reduced the elevation in lung water and protein permeability, while CGRP receptor blockade (CGRP8–37) had no effect (Fig. 5, A and B).

To further test whether PAR2-AP-induced lung inflammation and protein permeability were dependent on the NK1 and the NK2 receptor, we instilled PAR2-AP into lungs of mice pretreated with neuropeptide antagonists of NK1 and NK2 receptors (Spantide III + GR87389) or capsaicin and performed BAL at 4 h. The PAR2-AP-induced increases in protein concentration, number of neutrophils, and MIP-2 level in the BAL were blunted in mice pretreated with Spantide III plus GR87389 (NK1 + NK2 receptor blockade) or with capsaicin compared with vehicle-pretreated mice (Fig. 6, A–C). In addition, histological examination of lungs collected 4 h after PAR2-AP instillation revealed that the increase in marginated and emigrated neutrophils as well as protein-rich fluid in the airspaces was blunted in the capsaicin group compared with control (Fig. 6, D–G).

In this study, activation of PAR2 by intratracheal administration of PAR2-AP induced airway constriction, increased lung vascular and epithelial permeability to protein, and caused pulmonary edema in a dose-dependent manner. PAR2 activation also induced acidosis, hypoxemia, and hypercapnia as well as infiltration of leukocytes into lung parenchyma. These responses were not observed in the control peptide group and the PAR2 knockout mice, suggesting that they were indeed mediated by PAR2 and indicated a specific effect of the PAR2-AP. Airspace PAR2 activation also triggered substance P release and increased MIP-2 production. Capsaicin treatment, which ablates afferent sensory neurons, and pharmacological blockade of receptors for substance P (NK1) and neurokinin-A (NK2) markedly reduced these and other inflammatory responses to PAR2-AP. Our findings suggest that PAR2 activation can trigger a variety of inflammatory responses associated with lung injury and that these responses are almost entirely neuropeptide dependent.

Several of our observations are consistent with previous reports. The observation that instillation of PAR2-AP into the airway doubled the peak airway pressure and increased plateau pressure in wild-type but not PAR2 knockout mice is consistent with reports that PAR2 promotes hyperreactivity in allergic inflammation of the airway in mice (8), that PAR2-AP can increase the contractile response of isolated bronchi from guinea pigs to histamine (32), and that PAR2-AP can induce contraction of human airways and potentiate contractions to histamine (13, 14). Our observation that PAR2-AP can trigger leukocyte infiltration into lung is consistent with reports that PAR2-AP can promote leukocyte rolling and arrest on mesenteric (10) and cremasteric vessels (11) and neurogenic inflammation in rat paw (5). The lack of response in the control peptide group and in the PAR2 knockout mice in these studies supports the conclusion that most of proinflammatory responses to PAR2-AP are indeed PAR2 mediated.

The magnitude of PAR2-AP-induced responses in the lung was impressive. In addition to doubling airway pressure, PAR2-AP 5-fold increased substance P level in the BAL and nearly doubled the ELW in only 30 min, and by 4 h increased protein concentration in BAL by 50%, MIP-2 levels by 8-fold, and the number of neutrophils by 12-fold. Plasma MDA levels were increased 4-fold. These changes were functionally significant because they were associated with hypoxemia, hypercapnia, and acidosis. Thus, PAR2 activation in the lung can significantly impair ventilation and gas exchange.

In principle, the various inflammatory responses caused by PAR2-AP may be mediated by PAR2 expressed on any of a number of different cell types. Airway epithelial cells would be the first cells to encounter acting peptide after intratracheal administration, and PAR2 expression is increased in the airway epithelium of patients with allergic asthma (33). Activation of PAR2 has been reported to cause bronchorelaxation ex vivo in a manner that depends upon the presence of airway epithelium, perhaps through epithelial production of PGE₂ (12). We measured PGE₂ in the BAL 30 min after Hanks’, PAR2-AP, and LSIGRL instillation in this study; PAR2-AP triggered only a modest increase in the level of PGE2 (530 ± 10 pg/ml, n = 6) relative to Hanks’ (490 ± 40 pg/ml, n = 10) and LSIGRL (467 ± 71 pg/ml, n = 6). Thus, at least in our acute in vivo studies, the PG response was small and the effects of PAR2 activation on airway pressures suggest bronchoconstriction was predominant response.

PAR2 is also expressed on mouse microvascular endothelial cells (34), and PAR2 may be expressed by and contribute to a variety of functions in human neutrophils (35, 36). Thus, a direct effect of PAR2-AP on neutrophils might contribute to increased MPO activity and MDA levels in plasma, and perhaps additional responses in the lung in these studies. Intriguingly, however, capsaicin pretreatment of mice markedly reduced the ability of PAR2-AP to trigger increases in airway pressure, lung water and lung vascular permeability, protein content, MIP-2 production, and the number of neutrophils in the airspaces and interstitium. Antagonists of neuropeptide function also prevented these responses to PAR2-AP, and PAR2-AP triggered neuropeptide release as indicated by increased substance P levels in BAL. Overall, our results suggest that PAR2-AP-induced inflammatory responses in the lung are substantially dependent upon neurogenic mechanisms. Of note, however, inhibition of PAR2-AP-induced increases in airway pressure by capsaicin and neuropeptide antagonists was less complete than inhibition of other indices of inflammation (Figs. 4 and 6), suggesting that a neuropeptide-independent component, perhaps mediated by the airway epithelium, contributes to the former response.

Neurogenic mechanisms have been implicated in lung inflammation, and PAR2 has been implicated in neurogenic inflammation in other tissues. Airways are innervated by bronchopulmonary capsaicin-sensitive C-fiber afferents that are carried by the vagus nerve (37), and some spinal DRG neurons (38). Sensory DRG neurons coexpress substance P, CGRP, and PAR2 (5, 21), and neuropeptides in the airways can induce airway constriction (through NK2 receptor) and peripheral lung vascular leakage (NK1 receptor). Neuropeptides have been reported to contribute to lung inflammation associated with hepatic ischemia/reperfusion (39), and NK1 receptor activation can promote neutrophil recruitment to alveoli (38).

Based on the known effects of activating PAR1 on endothelial cells, we had expected that PAR peptides might act on endothelial cells directly to initiate edema formation and perhaps recruitment of neutrophils. Our results suggest that activation of PAR1 alone is not sufficient to trigger lung inflammation, at least under the conditions used in our studies. By contrast, these results suggest that activation of PAR2 alone is sufficient, and the neurogenic mechanism involved is consistent with a recent report that, unlike PAR1-AP, PAR2-AP is unable to directly enhance endothelial permeability in culture (40).

Epithelial and endothelial cells are both candidate targets of neurokinins released upon activation of PAR2 in the lung (17, 41, 42). Also, the alveolar macrophage is a potential effector cell. Alveolar macrophages are a source of proinflammatory cytokines in acute lung injury (43), and isolated mouse alveolar macrophages do express the NK1 receptor (44). Interaction of alveolar macrophages with NK1 released from subepithelial sensory neurons stretched by high tidal volume ventilation may help propagate lung inflammation (44). Thus, it is possible that the action of neuropeptides upon alveolar macrophages contributes in part to PAR2-AP-induced lung inflammation.

Taken together, our results indicate that PAR2 activation in lung can cause airway constriction, lung inflammation, and pulmonary edema, and that these responses are dependent upon, and perhaps mediated by, neuropeptides. PAR2 can be activated by a variety of extracellular proteases with trypsin-like specificity, including coagulation proteases (34, 45), tryptase (46), membrane-tethered serine protease-1 (47), and allergens (e.g., dust mite and cockroach proteases) (48). Which if any of these proteases activates PAR2 in vivo and the relative importance of PAR2 activation among potentially redundant signaling pathways that orchestrate inflammatory responses in the lung remain to be explored.

We thank Drs. Xiaohui Fang, Mark Looney, Yuanlin Song, and James Frank for their technical assistance and helpful suggestions.

The authors have no financial conflict of interest.