Search

Objective: We investigated and compared the in vivo effects of all four angiopoietins (COMP-Ang1, Ang2, Ang3, and Ang4) on blood and lymphatic vascular remodeling in adult mice. We analyzed the microvasculature of trachea and ear skin, and compared quiescent skin microvasculature with that during wound healing., Methods and Results: We were able to achieve similar levels of relatively long-term and sustained circulating expression of each angiopoietin using an adenoviral delivery system. Two weeks after treatment, we observed tracheal blood and lymphatic vascular enlargement, and lymphatic filopodia formation, with the following order of potency: COMP-Ang1>Ang3=Ang4>Ang2. Co-treatment with Ang2 attenuated Ang1-induced tracheal blood and lymphatic remodeling. In the normal ear skin, all angiopoietins induced blood vessel enlargement, whereas none induced lymphatic vascular remodeling. However, in the healing margin of ear skin wounds, all angiopoietins strongly induced lymphatic vascular enlargement and formation of lymphatic sprouts and filopodia, while they potentiated blood vascular enlargement. Co-treatment of Ang2 with Ang1 produced an additive effect on these changes., Conclusion: This study, one of the first to our knowledge to characterize the in vivo actions of all 4 angiopoietins, may expand the current concepts for use of angiopoietins for therapeutic angiogenesis and lymphangiogenesis.

VEGF, nitric oxide (NO), inflammation, and vascular- and extravascular remodeling coexist in asthma and other disorders. In these responses, VEGF regulates angiogenesis. VEGF also induces inflammation and remodeling. The mechanisms of the latter responses have not been defined, however. We hypothesized that VEGF-induces extravascular tissue responses via NO-dependent mechanisms. To evaluate this hypothesis, we compared the effects of transgenic VEGF165 in lungs from normal mice, mice treated with pan-NO synthase (NOS) or endothelial NOS (eNOS) inhibitors, and mice with null mutations of inducible NOS (iNOS) or eNOS. These studies demonstrate that VEGF selectively stimulates eNOS and iNOS. They also demonstrate that VEGF induces pulmonary alterations via NO-dependent and -independent mechanisms with angiogenesis, edema, mucus metaplasia, airway hyperresponsiveness, lymphocyte accumulation, dendritic cell hyperplasia and S-nitrosoglutathione reductase stimulation being NO-dependent and dendritic cell activation being NO-independent. Furthermore, they demonstrate that eNOS and iNOS both contribute to these responses. NO/NOS-based interventions may be therapeutic in VEGF-driven inflammation and remodeling.

Chronic inflammation in the airways is associated with dramatic architectural changes in the walls of the airways and in the vasculature they contain. In this study, we show that the adaptive immune system is essential for airway remodeling that occurs in mice that are chronically infected with the respiratory pathogen Mycoplasma pulmonis. Angiogenesis, lymphangiogenesis, and epithelial remodeling were greatly reduced in mice that lacked B cells. Substantiating a role for Ab and airway immune complexes, we found that the transfer of immune serum to B cell-deficient mice could reconstitute pathogen-induced angiogenesis. Inflammatory cells recruited to the infected airways were activated by the humoral response, and this activation correlated with the induction of genes for remodeling factors such as vascular endothelial growth factor-D. The results reveal a novel pathway whereby T cell-dependent humoral immunity to a persistent airway infection can induce inflammation-dependent angiogenesis, lymphangiogenesis, and chronic airway pathology.

Vascular enlargement is a characteristic feature of angiopoietin-1 (Ang1)-induced changes in adult blood vessels. However, it is unknown whether tissues having Ang1-mediated vascular enlargement have more blood flow or whether the enlargement is reversible. We have recently created a soluble, stable and potent Ang1 variant, COMP-Ang1. In the present study, we investigated the effects of varied dose and duration of COMP-Ang1 on vascular enlargement and blood flow in the tracheal microvasculature of adult mice and explored a possible mechanism of long-lasting vascular enlargement. We found that COMP-Ang1 administered by adenoviral vector induced long-lasting vascular enlargement and increased tracheal blood flow. In contrast, short-term administration of COMP-Ang1 recombinant protein induced transient vascular enlargement that spontaneously reversed within a month. In both cases, the vascular enlargement resulted from endothelial proliferation. The COMP-Ang1-induced vascular remodeling is mediated mainly through Tie2 activation. Sustained overexpression of Tie2 could participate in the maintenance of vascular changes. Together, our findings indicate that sustained treatment with COMP-Ang1 can produce long-lasting vascular enlargement and increased blood flow.

Integrin alpha5beta1 is among the proteins overexpressed on tumor vessels and is a potential target for diagnostics and therapeutics. Here, we mapped the distribution of alpha5beta1 integrin in three murine tumor models and identified sites of expression that are rapidly accessible to intravascular antibodies. When examined by conventional immunohistochemistry, alpha5beta1 integrin expression was strong on most blood vessels in RIP-Tag2 transgenic mouse tumors, adenomatous polyposis coli (apc) mouse adenomas, and implanted MCa-IV mammary carcinomas. Expression increased during malignant progression in RIP-Tag2 mice. However, immunoreactivity was also strong in normal pancreatic ducts, intestinal smooth muscle, and several other sites. To determine which sites of expression were rapidly accessible from the bloodstream, we intravenously injected anti-alpha5beta1 integrin antibody and 10 minutes to 24 hours later examined the amount and distribution of labeling. The injected antibody strongly labeled tumor vessels at all time points but did not label most normal blood vessels or gain access to pancreatic ducts or intestinal smooth muscle. Intense vascular labeling by anti-alpha5beta1 integrin antibody co-localized with the uniform CD31 immunoreactivity of tumor vessels and contrasted sharply with the patchy accumulation of nonspecific IgG at sites of leakage. This strategy of injecting antibodies revealed the uniform overexpression and rapid accessibility of alpha5beta1 integrin on tumor vessels and may prove useful in assessing other potential therapeutic targets in cancer.

Integrin alpha(5)beta(1) is overexpressed on endothelial cells of tumor vessels and is uniformly and rapidly accessible to antibodies in the bloodstream. Here, we determined whether antibodies rapidly gain access to integrin overexpressed on the abluminal (basolateral) surface of endothelial cells through vascular leakiness or whether the rapid accessibility results instead because the integrin is overexpressed on the luminal (apical) surface of endothelial cells due to loss of cell polarity. Using tumors in RIP-Tag2 transgenic mice as a model, we first compared the binding pattern of intravascular anti-alpha(5)beta(1) integrin antibody with the leakage pattern of nonspecific IgG. The distributions did not match: anti-alpha(5)beta(1) integrin antibody uniformly labeled the tumor vasculature, but IgG was located in patchy sites of leakage. We next injected an antibody to fibrinogen/fibrin, which resulted in patchy labeling of tumors that matched the leakage of IgG and the overall distribution of fibrin in tumors. Similarly, injected antibodies to the basement membrane protein fibronectin, a ligand of alpha(5)beta(1) integrin, or type IV collagen produced patchy sites of leakage instead of uniform labeling of vascular basement membrane. Differences in the kinetics of labeling, which for alpha(5)beta(1) integrin antibody was near maximal by 10 minutes but for the other antibodies gradually increased over 6 hours, indicated differences in accessibility of their respective targets. Isosurface rendering of confocal microscopic images was consistent with antibody binding to alpha(5)beta(1) integrin on the luminal surface of endothelial cells. Together, these findings indicate that the rapid accessibility of alpha(5)beta(1) integrin in RIP-Tag2 tumors results from overexpression of the integrin on the luminal surface of tumor vessels.

Edema occurs in asthma and other inflammatory diseases when the rate of plasma leakage from blood vessels exceeds the drainage through lymphatic vessels and other routes. It is unclear to what extent lymphatic vessels grow to compensate for increased leakage during inflammation and what drives the lymphangiogenesis that does occur. We addressed these issues in mouse models of (a) chronic respiratory tract infection with Mycoplasma pulmonis and (b) adenoviral transduction of airway epithelium with VEGF family growth factors. Blood vessel remodeling and lymphangiogenesis were both robust in infected airways. Inhibition of VEGFR-3 signaling completely prevented the growth of lymphatic vessels but not blood vessels. Lack of lymphatic growth exaggerated mucosal edema and reduced the hypertrophy of draining lymph nodes. Airway dendritic cells, macrophages, neutrophils, and epithelial cells expressed the VEGFR-3 ligands VEGF-C or VEGF-D. Adenoviral delivery of either VEGF-C or VEGF-D evoked lymphangiogenesis without angiogenesis, whereas adenoviral VEGF had the opposite effect. After antibiotic treatment of the infection, inflammation and remodeling of blood vessels quickly subsided, but lymphatic vessels persisted. Together, these findings suggest that when lymphangiogenesis is impaired, airway inflammation may lead to bronchial lymphedema and exaggerated airflow obstruction. Correction of defective lymphangiogenesis may benefit the treatment of asthma and other inflammatory airway diseases.

Tumor blood vessels have multiple structural and functional abnormalities. They are unusually dynamic, and naturally undergo sprouting, proliferation, remodeling or regression. The vessels are irregularly shaped, tortuous, and lack the normal hierarchical arrangement of arterioles, capillaries and venules. Endothelial cells in tumors have abnormalities in gene expression, require growth factors for survival and have defective barrier function to plasma proteins. Pericytes on tumor vessels are also abnormal. Aberrant endothelial cells and pericytes generate defective basement membrane. Angiogenesis inhibitors can stop the growth of tumor vessels, prune existing vessels and normalize surviving vessels. Loss of endothelial cells is not necessarily accompanied by simultaneous loss of pericytes and surrounding basement membrane, which together can then provide a scaffold for regrowth of tumor vessels. Rapid vascular regrowth reflects the ongoing drive for angiogenesis and bizarre microenvironment in tumors that promote vascular abnormalities and thereby create therapeutic targets.

Exaggerated levels of VEGF (vascular endothelial growth factor) are present in persons with asthma, but the role(s) of VEGF in normal and asthmatic lungs has not been defined. We generated lung-targeted VEGF(165) transgenic mice and evaluated the role of VEGF in T-helper type 2 cell (T(H)2)-mediated inflammation. In these mice, VEGF induced, through IL-13-dependent and -independent pathways, an asthma-like phenotype with inflammation, parenchymal and vascular remodeling, edema, mucus metaplasia, myocyte hyperplasia and airway hyper-responsiveness. VEGF also enhanced respiratory antigen sensitization and T(H)2 inflammation and increased the number of activated DC2 dendritic cells. In antigen-induced inflammation, VEGF was produced by epithelial cells and preferentially by T(H)2 versus T(H)1 cells. In this setting, it had a critical role in T(H)2 inflammation, cytokine production and physiologic dysregulation. Thus, VEGF is a mediator of vascular and extravascular remodeling and inflammation that enhances antigen sensitization and is crucial in adaptive T(H)2 inflammation. VEGF regulation may be therapeutic in asthma and other T(H)2 disorders.

Murine Mycoplasma pulmonis infection induces chronic lung and airway inflammation accompanied by profound and persistent microvascular remodeling in tracheobronchial mucosa. Because matrix metalloproteinase (MMP)-2 and -9 are important for angiogenesis associated with placental and long bone development and skin cancer, we hypothesized that they contribute to microvascular remodeling in airways infected with M. pulmonis. To test this hypothesis, we compared microvascular changes in airways after M. pulmonis infection of wild-type FVB/N mice with those of MMP-9(-/-) and MMP-2(-/-)/MMP-9(-/-) double-null mice and mice treated with the broad-spectrum MMP inhibitor AG3340 (Prinomastat). Using zymography and immunohistochemistry, we find that MMP-2 and MMP-9 rise strikingly in lungs and airways of infected wild-type FVB/N and C57BL/6 mice, with no zymographic activity or immunoreactivity in MMP-2(-/-)/MMP-9(-/-) animals. However, microvascular remodeling as assessed by Lycopersicon esculentum lectin staining of whole-mounted tracheae is as severe in infected MMP-9(-/-), MMP-2(-/-)/MMP-9(-/-) and AG3340-treated mice as in wild-type mice. Furthermore, all groups of infected mice develop similar inflammatory infiltrates and exhibit similar overall disease severity as indicated by decrease in body weight and increase in lung weight. Uninfected wild-type tracheae show negligible MMP-2 immunoreactivity, with scant MMP-9 immunoreactivity in and around growing cartilage. By contrast, MMP-2 appears in epithelial cells of infected, wild-type tracheae, and MMP-9 localizes to a large population of infiltrating leukocytes. We conclude that despite major increases in expression, MMP-2 and MMP-9 are not essential for microvascular remodeling in M. pulmonis-induced chronic airway inflammation.

Profiling gene expression in endothelial cells advances the understanding of normal vascular physiology and disease processes involving angiogenesis. However, endothelial cell purification has been challenging because of the difficulty of isolating cells and their low abundance. Here we examine gene expression in endothelial cells freshly isolated from lung capillaries after in vivo labeling with fluorescent cationic liposomes and purification by fluorescence-activated cell sorting (FACS). Of the 39,000 genes and expressed sequence tags evaluated on custom oligonucleotide arrays, 555 were enriched in endothelial cell fraction. These included familiar endothelial cell-associated genes such as VEGF, VEGF receptor (VEGFR)-1, VEGFR-2, angiopoietin-2, Tie1, Tie2, Edg1 receptor, VE-cadherin, claudin 5, connexin37, CD31, and CD34. Also enriched were genes in semaphorin/neuropilin (Sema3c and Nrp1), ephrin/Eph (ephrin A1, B1, B2, and EphB4), delta/notch (Hey1, Jagged 2, Notch 1, Notch 4, Numb, and Siah1b), and Wingless (Frizzled-4 and Tle1) signaling pathways involved in vascular development and angiogenesis. Expression of representative genes in alveolar capillary endothelial cells was verified by immunohistochemistry. Such expression reflects features that endothelial cells of normal lung capillaries have in common with embryonic and growing blood vessels. About half of the enriched genes, including exostosin 2, lipocalin 7, phospholipid scramblase 2, pleckstrin 2, protocadherin 1, Ryk, scube 1, serpinh1, SNF-related kinase, and several tetraspanins, had little or no previous association with endothelial cells. This approach can readily be used to profile genes expressed in blood vessels in tumors, chronic inflammation, and other sites in which endothelial cells avidly take up cationic liposomes.

Angiogenesis and microvascular remodeling are features of chronic airway inflammation caused by Mycoplasma pulmonis infection in rats. As airway blood vessels undergo remodeling, they become unusually sensitive to substance P-induced plasma leakage. Here we determined whether the remodeled vessels are leaky under baseline conditions, whether their heightened sensitivity is specific to substance P, and whether the leakage is reversible. Four weeks after infection, the amount of baseline leakage of Evans blue in the tracheal mucosa was two to five times the normal level. Gaps < 1 microm in diameter were located between endothelial cells in some remodeled vessels. Substance P, but not platelet-activating factor or 5-hydroxytryptamine, produced an exaggerated leakage response. Inhalation of the beta2-adrenergic receptor agonist salmeterol reduced the leakage by <60%. We conclude that the blood vessel remodeling after M. pulmonis infection is associated with microvascular leakiness due, in part, to the formation of endothelial gaps. This leakage is accompanied by an abnormal sensitivity to substance P but not to platelet-activating factor or 5-hydroxytryptamine and can be reduced by beta2-agonists.

Dendritic cells are antigen-presenting cells that constitutively express high levels of major histocompatibility complex class II (Ia) antigen on their plasma membrane. Previous studies have shown that the number of dendritic cells in the rat airway mucosa decreases rapidly after glucocorticoid treatment. We sought to determine whether apoptosis contributes to this steroid-induced cell decrease. Dendritic cells in tracheal whole mounts were revealed by immunoperoxidase staining using the OX-6 (anti-Ia) monoclonal antibody. In untreated rats, a dense network of Ia-immunoreactive (Ia+) cells with highly branched cytoplasmic processes was observed just beneath the tracheal epithelium (1,405 +/- 140 cells/mm2 mucosa; mean +/- SEM, n = 6). In rats treated with dexamethasone (10 mg/kg, intraperitoneally), four distinct changes in dendritic cell morphology were evident 4 to 8 h after injection: (1) appearance of large Ia+ granules in cytoplasmic processes, (2) narrowing of cytoplasmic processes, (3) loss of Ia immunoreactivity from the cell surface, and (4) fragmentation of cells into small Ia+ bodies. These changes accompanied a 56% decrease in the number of Ia+ cells over 8 h. The contribution of apoptosis to this decrease in Ia+ cells was determined by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end-labeling (TUNEL) of nucleosomal DNA fragments in histologic sections. The number of TUNEL+ bodies increased from a control value of 174 +/- 47 bodies/mm2 mucosa to 2,108 +/- 294 bodies/mm2 mucosa at 4 h and 936 +/- 343 bodies/ mm2 mucosa at 8 h (n = 4 rats per time point). The location of TUNEL+ bodies closely corresponded to that of Ia+ cells stained in adjacent histologic sections. We conclude that apoptosis contributes to the rapid decrease in airway dendritic cells after glucocorticoid treatment.

Exposure of sensitized individuals to antigen can induce allergic responses in the respiratory tract, manifested by early and late phases of vasodilatation, plasma leakage, leukocyte influx, and bronchoconstriction. Similar responses can occur in the skin, eye, and gastrointestinal tract. The early-phase response involves mast cell mediators and the late-phase response is leukocyte dependent, but the mechanism of leakage is not understood. We sought to identify the leaky blood vessels, to determine whether these vessels contained endothelial gaps, and to analyze the relationship of the gaps to adherent leukocytes, using biotinylated lectins or silver nitrate to stain the cells in situ and Monastral blue as a tracer to quantify plasma leakage. Most of the leakage occurred in postcapillary venules (< 40-microns diameter), whereas most of the leukocyte migration (predominantly neutrophils) occurred in collecting venules. Capillaries and arterioles did not leak. Endothelial gaps were found in the leaky venules, both by silver nitrate staining and by scanning electron microscopy, and 94% of the gaps were distinct from sites of leukocyte adhesion or migration. We conclude that endothelial gaps contribute to both early and late phases of plasma leakage induced by antigen, but most leakage occurs upstream to sites of leukocyte adhesion.

Neurogenic inflammation, in its original definition, the plasma leakage induced by stimulation of peripheral sensory nerves, occurs in the postcapillary venules of the skin and airways. Plasma leakage is accompanied by increased blood flow, which results from dilatation of arterioles. In skin, these phenomena are manifested as wheal and flare, respectively. Both phenomena are mediated by neuropeptides released from capsaicin-sensitive unmyelinated sensory nerve fibers. Substance P is the primary mediator responsible for plasma leakage, acting via tachykinin NK-1 receptors, whereas both calcitonin gene-related peptide and substance P induce vasodilatation. Sensory nerve transmitters also cause release of histamine from mast cells, which contributes substantially to plasma leakage in the skin, but less so in the airways. Substance P causes an increase in vascular permeability as a result of the focal, transient, and fully reversible formation of gaps, approximately 0.5 to 1.5 microns in diameter, located in the intercellular junctions of endothelial cells. The gaps can be visualized by silver nitrate staining of the endothelial cell borders, by lectin staining, or by scanning and transmission electron microscopy. Neurogenic inflammation can be inhibited by preventing the stimulation of sensory nerves, by presynaptic inhibition of neuropeptide release from sensory nerves, or by blocking neuropeptide receptors. The formation of endothelial gaps can also be inhibited by anti-inflammatory drugs that stabilize endothelial cells, such as beta-adrenergic agonists and steroids.

Understanding the physiological role of tachykinins requires precise cellular and subcellular localization of their receptors. We raised antisera by immunizing rabbits with peptides corresponding to portions of the intracellular tails of the rat neurokinin 1, 2, and 3 receptors (NK1-R, NK2-R, NK3-R). Receptors were localized by immunofluorescence and confocal microscopy. NK1-R, NK2-R, and NK3-R were detected at the plasma membrane of transfected cells with minimal intracellular stores. Staining was abolished by preabsorption of the antisera with the peptides used for immunization. Nontransfected cells were unstained. Each antiserum only stained cells transfected with the appropriate receptor and did not stain cells transfected with the other receptors. Therefore, the antisera are specific and do not cross-react with other neurokinin receptors. We examined the distribution of the neurokinin receptors in the gastrointestinal tract of the rat. NK1-R was detected in myenteric and submucosal neurons and in interstitial cells of Cajal. NK2-R was localized to circular and longitudinal muscle cells and to nerve endings in the plexuses. NK3-R was detected in numerous myenteric and submucosal neurons. Some neurons expressed both NK1-R and NK3-R. Receptors were detected at the plasma membrane and in endosomes. Cells expressing the receptors were closely associated with tachykinin-containing nerve fibers. Thus, NK1-R and NK3-R mediate neurotransmission by tachykinins within enteric nerve plexuses, and NK1-R and NK2-R mediate the effects of tachykinins on interstitial and smooth muscle cells, respectively.

Several lines of evidence suggest that sensory nerves of the carotid body have an efferent function in addition to their afferent function of conducting chemoreceptive impulses to the brain. However, it has been difficult to document the release of substances from sensory nerve terminals on glomus cells and to determine whether such an efferent function plays a role in chemoreception. By comparison, the phenomenon of neurogenic inflammation has been relatively easy to study in rats and guinea pigs and has proven to be an informative model system for analyzing efferent actions of sensory nerves. The main characteristic of neurogenic inflammation is plasma leakage. Chemical irritants that activate unmyelinated sensory nerves cause plasma leakage in the skin, respiratory tract, and other organs by triggering the release of substances from sensory nerve fibers. Substance P, which is synthesized and released by some sensory neurons, appears to be the main active mediator, although other tachykinins, calcitonin gene-related peptide, and perhaps other peptides may also participate. Neurogenic inflammation results from the action of substance P on NK1 receptors, as demonstrated by selective NK1 receptor agonists and antagonists. The NK1 receptors involved in plasma leakage are located on the endothelial cells of postcapillary venules and collecting venules. Within seconds of the activation of NK1 receptors by substance P, gaps form in the endothelium of target vessels. The endothelial gaps are transient, and the leak normally ends in a few minutes. However, the magnitude of the response can increase in pathological conditions such as Mycoplasma pulmonis infection in rats, which results in a chronic inflammatory disease of the respiratory tract. The infected airway mucosa becomes abnormally vascular as a result of angiogenesis, and the endothelial cells of the newly formed vessels express increased numbers of NK1 receptors and thus are abnormally sensitive to substance P. Studies of neurogenic inflammation not only have helped to understand the efferent actions of sensory nerves but also have given insight into the mechanism and consequences of inflammatory changes in endothelial cells and in the plasma leakage that follows.

The tracheal epithelium of pathogen-free rats consists mainly of serous-type secretory cells, ciliated cells, basal cells, and a few neuroendocrine cells. Mucus-containing goblet cells are rare. Calcitonin gene-related peptide (CGRP) is known to exist in the neuroendocrine cells and in sensory nerves of the tracheal mucosa and is released into the airway lumen by sensory nerve stimulation. In this study, we determined whether epithelial serous cells are another source of CGRP. Tracheas of adult male specific pathogen-free F344 rats were immunostained by an avidin-biotin technique either as whole mounts or as cryostat sections using two different polyclonal primary antibodies to rat CGRP. Some specimens were stained for CGRP-like immunofluorescence and examined with a confocal microscope. CGRP immunoreactivity was present in granules of serous cells throughout the trachea. In whole mounts, the stained cells were most abundant between the cartilaginous rings, especially in the rostral trachea, where they constituted 56% of the epithelial cells in contact with the tracheal lumen. Serous cells were easily distinguished from neuroendocrine cells and nerve fibers with CGRP immunoreactivity. In evidence that the CGRP immunoreactivity was specific, the staining of serous cells was abolished by omitting the primary antibody and by absorption with 10 micrograms/ml CGRP. Antibodies to substance P, vasoactive intestinal polypeptide, and tyrosine hydroxylase did not stain epithelial serous cells. An antibody to protein gene product 9.5 labeled neuroendocrine cells, but not serous cells. Injection of capsaicin (150 micrograms/kg intravenously), a substance known to degranulate epithelial serous cells, reduced the staining of the serous cells for CGRP.(ABSTRACT TRUNCATED AT 250 WORDS)

Substance P is one of the peptides released from sensory nerves that mediate "neurogenic inflammation." Although substance P-immunoreactive (SP-IR) axons are known to be present within the mucosa of the respiratory tract, the relative extent of the innervation of various components of the mucosa is not known. Therefore, we determined the distribution and number of SP-IR axons in the rat trachea and bronchi, by using immunohistochemistry on tissue whole mounts. Specifically, we sought to learn whether these axons directly innervate the postcapillary venules involved in neurogenic plasma extravasation, the arterioles involved in neurogenic vasodilatation, and the airway smooth muscle involved in bronchoconstriction in pathogen-free, adult male F344 rats. We found that 90% of the SP-IR axons were single axons, usually having varicosities. Eighty-five percent of these were in the epithelium, 6% innervated arterioles, and the remainder elsewhere in the lamina propria. Only 10% of the mediator-sensitive postcapillary venules (i.e., venules labeled with Monastral blue pigment after challenge with capsaicin or substance P) were within 10 microns of SP-IR axons. SP-IR axons were more than 10 times as frequent in the smooth muscle of the distal bronchi as in the trachea. Capsaicin pretreatment (168 mg/kg over 7 days) reduced the number of SP-IR axons in the trachea by 96%, which is consistent with their being sensory. Unilateral vagotomy reduced the number of SP-IR axons bilaterally in the trachea and ipsilaterally in the main stem bronchus. Using an antibody to Protein Gene Product 9.5 as a nonspecific marker for all nerves in the trachea, we determined that SP-IR axons constituted 90% of the axons in the epithelium, 32% of the axons on arterioles, and only 4% of the axons in the smooth muscle. We conclude that most SP-IR nerves in the trachea are sensory axons and most of these axons end in the epithelium. SP-IR axons innervate mucosal arterioles, but few innervate postcapillary venules. Therefore, the mechanism by which sensory axons evoke plasma extravasation from these venules is likely to involve the diffusion of the peptide or a secondary mediator from the epithelium or from the arterioles upstream.