© 2002 American Thoracic Society
Three-Dimensional Mapping of Sensory Innervation with Substance P in Porcine Bronchial MucosaComparison with Human AirwaysDepartment of Physiology; and The Asthma Institute of Asthma and Allergy, Department of Medicine, University of Western Australia, Nedlands, Western Australia Correspondence and requests for reprints should be addressed to Jasmine P. Lamb, Department of Physiology, University of Western Australia, 35 Stirling Highway, Nedlands, Western Australia 6009. E-mail: jlamb{at}cyllene.uwa.edu.au
In asthma, neurogenic inflammation in bronchial airways may occur though the release of neuropeptides from C fibers via an axon reflex. Structural evidence for this neural pathway was sought in the pig and in humans by three-dimensional mapping of substance P–immunoreactive (SP-IR) nerves in whole mounts of mucosa using immunofluorescent staining and confocal microscopy. To show continuity, nerves were traced with 1,1'-didodecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate from their epithelial endings through the mucosa. The pan-neuronal marker protein gene product 9.5 revealed an extensive apical and basal plexus of nerves in the epithelium; 94% of these were varicose SP-IR fibers. Apical SP-IR fibers had a length density of 88 mm/mm2. Varicose apical processes followed closely around the circumference of goblet cells. Calcitonin gene–related peptide was colocalized with SP-IR in varicosites. The epithelial fibers converged into bundles as they entered the lamina propria where lateral branches ran along arterioles, often contiguous with the vascular smooth muscle. 1,1'-didodecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate tracing showed that they projected to the epithelium. SP-IR fibers were rare near postcapillary venules. In human bronchial epithelium, protein gene product 9.5 revealed a similar apical and basal plexus of varicose fibers that weakly stained for SP-IR. Thus, a continuous sensory nerve pathway from the epithelium to arterioles provides structural support for a local axon reflex.
Key Words: neurogenic inflammation • substance P • afferent pathways • bronchi/innervation • human/physiology • swine/physiology • neuropeptides
The sensory innervation of the bronchial mucosa comprises mainly varicose, unmyelinated, nociceptive C fibers that detect chemical substances present in the airway lumen (1). Sensory nerve endings in the epithelium respond to stimulation by generating impulses that are conducted centrally, and locally with the release of the proinflammatory neuropeptides, calcitonin gene-related peptide (CGRP), substance P (SP), and neurokinin A (2, 3). In the airway mucosa, neuropeptides cause microvascular plasma leakage (4–6), arteriolar dilation (7), mucous secretion (8), and recruitment of inflammatory cells (9, 10), collectively described as neurogenic inflammation. There is functional evidence that part of the neurogenic response, for example, plasma extravasation in rats (11) and bronchial vasodilatation in pigs (12) and dogs (13), can be attributed to a local axon reflex in tracheal and bronchial mucosa. This suggests that a sensory nerve pathway extends from nerve endings in the epithelium to underlying target vessels and glands where neuropeptides are released. This mechanism has been proposed as a cause of human asthma (14, 15), a condition in which increased levels of tachykinins and albumin in bronchial lavage fluid are found, indicating that tachykinin release and vascular leakage occurs (2). SP-immunoreactivity (SP-IR) and CGRP-IR nerves are found in the mucosa of airways in humans (16, 17), pigs (18), cats (16, 19), dogs (20), guinea pigs (16, 21, 22), and rats (4, 16) by immunohistochemistry. There is an abundance of SP-containing nerves in the bronchial epithelium in rats and guinea pigs (4, 16), but the occurrence of these nerves in larger mammals and in humans is controversial. Guinea pig and human mucosa have been reported to contain similar amounts of SP determined by radioimmunoassay (6, 16), whereas SP-IR nerves determined immunohistochemically were sparse in large mammals, including pigs and humans (23). Furthermore, although Lundberg and colleagues (16) and Ollerenshaw and colleagues (17) showed SP in the human bronchial epithelium, many later studies have found little or no SP-IR nerves (24–28). There remains strong functional evidence that in large mammals the bronchi are innervated by neuropeptide-containing sensory nerves. In response to capsaicin, a vasodilator response is mediated locally via an axon reflex and centrally via vagal cholinergic and noncholinergic parasympathetic nerves in pigs (12), dogs (13, 20), and cats (29). This suggests that the absence of SP, CGRP, or neurokinin A fibers mentioned previously here may have been methodological. A delay in fixation may lead to SP degradation by endogenous neutral endopeptidase. SP levels may decline with age (26) and exposure to irritants such as cigarette smoke (30). Conventional histology, which uses thin sections, is not suited for following nerve pathways with a complex three-dimensional arrangement. Neurogenic inflammation involving C fiber–mediated reflexes continues to be viewed as a cause of bronchial hyper-responsiveness and asthma (3, 10, 14, 31). With respect to local reflexes, it is hypothesized that there are continuous nerves pathways between the sensory nerve endings in the epithelium and mucosal blood vessels. To provide structural evidence for this mechanism, these fibers in the bronchial mucosa in humans and large animals need to be mapped. The aim of this study was threefold: first to characterize the distribution of SP- and CGRP-containing nerves in the young pig where loss of sensory nerves in the epithelium through age or environmental influences would be minimal. Pig lung tissues were therefore fixed rapidly in situ by perfusion via the trachea to avoid the breakdown of tachykinins by endopeptidases. The second aim, in the human bronchial mucosa, was to show that morphologically similar nerves were present in the epithelium by their immunoreactivity to protein gene product 9.5 (PGP-IR), a protein found in all nerve cells. The third aim was to demonstrate that the projections of the epithelial nerves to the blood vessels in the lamina propria are continuous using 1,1'-didodecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate (DiI), a lipophilic carbocyanine tracer (32). This study uses a new approach: tracing nerve pathways in three dimensions in whole mounts of mucosa from the epithelial surface to their target tissues in the lamina propria using the confocal microscope.
Porcine Bronchial Mucosa Landrace pigs (n = 6, 4 weeks old and 9–11 kg in body weight) were obtained from Medina Agricultural Farm (Medina, WA, Australia). Pigs were sedated with tiletamine/zolazepam (0.4 mg/kg intramuscularly; Virbac, Peakhurst, NSW, Australia) and xylazine (0.2 mg/kg; Troy Laboratories, Smithfield, NSW, Australia) and euthanized with sodium pentabarbitone (163 mg/kg intravenously). Within 5 minutes, an endotracheal tube (6.0 mm inside diameter) was installed into the distal trachea, and the lungs were deflated by carefully puncturing the chest wall and reinflated in situ to 15–20 cm H2O with fixative 4% paraformaldehyde in phosphate-buffered saline (pH 7.2) via the tracheal tube. Constant pressure was maintained for 1 hour. The trachea was ligated, and the lung was excised and immersed in fixative. The main stem bronchi were dissected out between second and fifth laterals (33). The internal diameter of each bronchus ranged from approximately 2–8 mm. The bronchus was opened out and pinned epithelial side up onto a petri dish lined with Sylgard (Dow Corning, Midland, MI). The mucosa was separated at the lamina propria/smooth muscle junction by peeling it away from the wall. Procedures were approved by the Animal Ethics Committee of the University of Western Australia.
Human Bronchial Mucosa
Tracing of Neural Pathways
Imunohistochemistry
Confocal Laser Scanning Microscopy The confocal microscope used was MRC1024 with Laser Sharp acquisition version 3.2 software (BIORAD; Hemel Hempsted, Hertfordshire, UK) and a Diaphot inverted microscope (Nikon, Japan) focusing a Krypron/Argon laser (American Laser Corp., Salt Lake City, UT). The excitation wavelengths of the laser were 488, 568, and 647 nm. Fluorophores Oregon Green 488, CY3, DiI, CY5, and Alexa Fluor 633 (Table 1) had absorption (excitation) maxima at 496, 554, 551, 649, and 632 nm, and their emission maxima were 524, 566, 585, 666, and 647 nm, respectively. In double- and triple-labeling experiments, the fluorophores were selected for maximal separation, and no bleed through occurred. Each whole mount was scanned at increasing depths of focus (0.5-µm steps) producing a z series of xy images. From this z series the brightest pixel at each xy location was projected to give a two-dimensional image using Confocal Assistant 4.2. The z series was also viewed in cross-section with National Institutes of Health image software. Cross-sections (x series of zy images and y series of zx images) were reconstructed and projected. Adjacent fields were montaged using Adobe Photoshop 5.0 software. In multiple-labeling experiments, single fields were scanned for each marker, and the images were colored and superimposed with Adobe Photoshop 5.0 software.
Nerves in the Epithelium In the epithelium, the length of the SP-IR nerves in the apical and basal layers was compared for each of six fields (three pigs). This was done with reference to three-dimensional images created from their z series. In the apical layer, the fibers were traced from their nerve endings to the point at which the axon began to descend through the epithelium (denoted "X"). The basal layer was traced from X to the point at which nerves descended into the lamina propria. These were measurements of horizontal length; it was not feasible to measure the vertical components. The nerve tracings were measured with National Institutes of Health image software to calculate the total length per unit area and were expressed as mean ± SEM. The number of nerve processes in the apical layer and the number of fibers rising up from the basal layer were also counted. The size of each field was 31,605 µm2.
Mapping of Sensory Nerves in Porcine Bronchial Mucosa An extensive innervation of the bronchial mucosa was demonstrated by immunofluorescence labeling of the nerves for PGP 9.5, a pan-neuronal marker, and confocal imaging to depths of 100 µm through whole mounts of mucosa. All images shown are orientated with the mucosal folds running from left to right. Figure 1A is a representative field of PGP-IR nerves imaged from the luminal surface. Tissue autofluorescence resulting from 488-nm excitation faintly reveals the epithelial cells enabling the depth of the epithelium to be determined during scanning (approximately 20 µm), and this allowed the distribution of nerves to be partitioned above and below the basal lamina. Figure 1B shows an abundance of varicose fibers in the epithelium, with fewer nerves in the trough of the mucosal fold. The lamina propria contained many nerve bundles that predominantly ran along the length of the airways (Figure 1C).
Specific staining for SP-IR demonstrated that the mucosa was innervated by fine fibers and a few nerve bundles (Figure 2A) . In the epithelium, these nerves had characteristically fine, varicose axons (approximately 1 µm in diameter), and their endings often had one to five large terminal varicosities (approximately 3 µm) (Figure 2B). For each of 20 fields (four pigs), the three-dimensional information contained in the z series was used to show the arrangement of SP-IR nerves in cross-section by digital reconstruction. Figure 3 shows a lumen view (large panel) together with projected cross-sections (right and bottom panels). These cross-sections showed SP-IR fibers lying in two plexi in the epithelium. An apical layer of SP-IR nerves lays just beneath the luminal border (0–6 µm) comprised of many fibers and nerve endings and is shown as a line diagram (Figure 3B) where X marks the point at which fibers leave this layer. These axons descended to the base of the epithelium, where they ran laterally forming a second plexus, also shown as a tracing (Figure 3C). In the apical layer, there were 1,904 ± 303 terminal processes per mm2, and the apical network was supplied by 709 ± 155 ascending nerves per mm2, which arose from a near continuous basal nerve network, that is, approximately three apical processes per (ascending) axon. The length of the processes in the apical layer ranged from 12 to 105 µm in Figure 3, and the maximum length was 126 µm for six fields from three pigs. The length–density of fibers was approximately fourfold greater in the apical layer compared with the basal layer, 87.5 ± 5.15 and 21 ± 6 mm/mm2, respectively. We did not attempt to measure the vertical component of fibers passing through the epithelium to the basal layer.
To determine the population of nerves in the epithelium that contain SP, the mucosa was stained for PGP and SP concurrently (Figure 4) . In seven fields from three pigs, 94% of the PGP-IR nerves were SP-IR (significant, p < 0.05).
Fluorescently labeled phalloidin toxin binding to F-actin outlined the cells in the apical epithelium, enabling unstained goblet cells to be distinguished among the epithelial cells (Figure 5A) . At high power, tufts of cilia at their luminal border were observed (Figure 5Ai). In this layer, epithelial nerves, both SP-IR and PGP-IR positive, frequently passed around the apex of goblet cells (Figures 5B and 5C). Epithelial cells were visualized by their autofluorescence when excited by the 488-nm laser used to reveal PGP-IR (Figures 5C and 5D), whereas goblet cells appeared as black circles. Fibers passed between epithelial cells (Figure 5D). SP-IR fibers in the basal epithelium converged into larger bundles as they penetrated the basement membrane and entered the lamina propria (Figure 6) . At 20–30 µm into the lamina propria, these bundles ran longitudinally along the airway, whereas other bundles continued to penetrate deeper into the lamina propria. In Figure 6, nerves deep in the lamina propria were faint because the laser penetration was at its limit. In subsequent preparations, this was overcome by separating the mucosa just below these nerve trunks and scanning from the lamina propria side of the whole mount. Figure 7 shows large SP-IR nerve bundles in the lamina propria from which smaller bundles branched upward to give rise to the epithelial nerves. These nerves ran longitudinally along the airways together with other large bundles revealed by PGP-IR (Figures 8A and 1C) . These SP-IR nerves retained their distinctive varicose appearance. At intervals along the airway some bundles passed through the muscularis to the adventitia (data not shown).
In the lamina propria, SP-IR fibers occurred in the proximity of the mucosal circulation and mucous glands (Figures 8B–8D). Antibodies against smooth muscle  -actin stained the muscularis and pericytes of blood vessels and the myoepithelial cells of glands. SP-IR nerves followed small arterioles, located approximately 30 µm below the basement membrane, that had closely packed smooth muscle bundles encircling the vessel. These small bundles followed the contours of the outer surface of the vascular smooth muscle at the same optical depth, indicating that they lay beside the muscularis (Figure 8B). Occasionally SP-IR fibers were seen to cross (less than 5 µm) precapillary arterioles or the capillary bed underneath the epithelium. SP-IR fibers were seldom seen within 10 µm of the postcapillary venules identified by -actin staining of the pericytes that surround the vessel, incompletely leaving claw–like spaces (Figure 8B). Larger bundles of SP-IR nerves ran along arterioles at approximately 10 µm (Figure 8C), and these were on occasions followed to mucous glands. SP-IR fibers were observed within mucous glands that had cells with unstained, cytoplasmic regions (Figure 8D). Whole mounts of mucosa stained for CGRP-IR (Figure 9A) had nerves with the same appearance and distribution as SP-IR (Figure 6). Concurrent staining for SP-IR and CGRP-IR demonstrated colocalization (Figures 9B and 9C).
Mapping of Sensory Nerves in Human Bronchial Mucosa There was a marked difference in the amount of PGP-IR nerves in the 39-year-old patient compared with the elderly patients. The former patient had an abundance of PGP-IR varicose fibers in the epithelium with larger terminal varicosities (Figure 10A) . A view of the whole mount in cross-section was obtained by digital reconstruction of the confocal z series (Figure 10A) and also by directly imaging the whole mount from the side (Figure 10B). Together, these views demonstrated the three-dimensional arrangement of the nerves lying in apical and basal plexi in the epithelium with nerves traversing between these layers. The apical layer of fibers with nerve endings lies just beneath the luminal border with nerves following around the apex of the goblet cells (Figures 10C–10E). The basal nerves converged into bundles that entered the lamina propria. Many nerve bundles were seen deep in the lamina propria by imaging from the muscularis side of the lamina propria (data not shown). In elderly patients, few PGP-IR epithelial nerves were seen. These occurred in small infrequent patches. Some profiles were seen ascending toward the apical surface in cryosections, but there were lengths of epithelium in which no fibers were observed.
Compared with PGP-IR staining of epithelial nerves, SP-IR was very faint and required greater laser excitation resulting in a high background. Figure 11A shows strings of varicosities detected in the epithelium of the 39-year-old patient. Some colocalization of PGP-IR and SP-IR was seen in epithelial nerves where SP-IR was detected (Figures 11B and 11C).
Tracing Neural Pathways in Porcine Bronchial Mucosa To demonstrate the continuity of the pathways of sensory nerves in the bronchial mucosa of the pig, DiI tracing and CGRP immunostaining were used to follow and identify these fibers, respectively. DiI was deposited on to the luminal surface of segments of pig bronchus. There it entered the endings of varicose nerve fibers in the apical layer of the epithelium and revealed a network of fine fibers (0.5–1.8 µm) that had enlarged terminal varicosities (2–2.5 µm) (Figure 12A) . Receptive fields, comprising apical epithelial fibers stemming from a common nerve, were distinguished with DiI. Figure 12B shows a typical receptive field (maximum span 89 µm) with five or more processes, and Figure 12C is a corresponding side view (a projection of y–z cross-sections) showing these fibers all originated from a bundle that had ascended from the lamina propria. DiI diffusion into nerve bundles and trunks in the lamina propria demonstrated continuity with fine epithelial fibers (Figure 12D). These trunks in the lamina propria also gave rise to nerve bundles running longitudinally along the airway (approximately 45 µm below the lumen). DiI-labeled trunks were comparable in diameter to those seen with PGP-IR (Figure 8A) and were much larger than SP-IR bundles observed (Figures 7 and 8A). This indicates that DiI diffused across the membranes of individual fibers into adjacent nerves. The dye also diffused into unidentified cells that must be in close contact with the nerve fibers. Two principle shaped cells were noted: round (Figures 12A) and star-shaped with long processes (data not shown).
Varicose epithelial nerves traced with DiI also stained positively for CGRP-IR (Figure 13A) . Colocalization of these labels confirmed that the DiI was mainly tracing sensory nerves in the epithelium. Immunostaining was weaker in glycerol-permeablized tissues compared with dimethyl sulfoxide treatment (Figure 9); however, this compromise was unavoidable as dimethyl sulfoxide dissolves DiI. The intensity of the CGRP-IR was strongest in the varicosities and was variable where DiI had previously diffused along the nerve. As a result, continuity of these strings of varicosities was often masked in projections. These nerve pathways were more readily followed by examining the optical sections of the z series.
Continuity between epithelial nerves and axons that run along blood vessels in the lamina propria was demonstrated by merging DiI tracings of nerves through the mucosa and the corresponding -actin-IR that visualized blood vessels. Figure 13B shows a DiI-labeled nerve in the lamina propria that had been traced from the epithelium (data not shown). It branches into two nerves that run longitudinally along the airway on either side of an arteriole. Concurrent CGRP immunostaining identified sensory fibers within DiI-labeled bundles with continuous nerve pathways from the epithelium to blood vessels. Figure 13C is a view of the lamina propria that shows two DiI/CGRP-IR–labeled bundles that branch from a nerve trunk and follow an arteriole. Another example of these pathways is shown through the whole mucosa in Figure 13D, where the field corresponds to Figure 12D. As a consequence of the thickness of this projection (82 optical sections), the appearance of the finer CGRP-IR nerves was poor for the reasons given previously here (Figure 13Di). In this large montage, CGRP-IR/DiI–labeled epithelial fibers join bundles that enter the lamina propria and pass through large nerve trunks into smaller bundles that follow blood vessels running longitudinally along the airway (Figure 13Dii). In most tissue samples, sensory nerves appeared predominantly to follow arterioles (Figures 13B and 13C) and rarely venules and post capillary venules (Figures 13Dii and 13Diii). These bundles often touched the vessels or followed them closely (less than 1–10 µm apart). Bundles labeled by DiI were often wider than the CGRP-IR stain, particularly in large nerve trunks consistent with Figure 8A where PGP-IR demonstrates a substantial number of the fibers are not SP-IR.
In this study, we have mapped the distribution of SP-IR and CGRP-IR nerves in whole mounts of porcine bronchial mucosa by imaging with confocal microscopy through depths of 100 µm. A comprehensive view of the three-dimensional organization in the mucosa and their anatomical relationship to the bronchial circulation and mucous-secreting cells has been obtained in a way that has not been previously possible using thin sections. It can be reasonably concluded that these varicose fibers are sensory nerves based on their distribution and neuropeptide immunoreactivity. In the lamina propria, varicose SP-IR and CGRP-IR nerves lie in nerve trunks adjacent to the airway smooth muscle and give rise to bundles that travel toward the epithelium. These nerves have lateral branches to within approximately 20 µm of the epithelium that run along the arterioles and go to mucous glands; smaller bundles are contiguous with the vascular smooth muscle of smaller arterioles. The nerves that reach the epithelium pass through the basement membrane and spread laterally to form a network at the base of the epithelium. From this, plexus fibers ascend between epithelial cells to the apex where they arborize to form a layer of varicose processes within 3 µm of the luminal surface. Many processes follow very closely around the circumference of goblet cells. The apical layer of epithelial nerves was the most densely innervated region of the mucosa and comprised single, varicose, SP-IR, and CGRP-IR processes spanning up to 126 µm that terminated in one or more enlarged varicosities. These appeared to be arborizations originating from a central axon having one to five terminal processes, confirmed by DiI tracing, which spread out laterally, forming irregularly shaped receptive fields that were elongated in the direction of the mucosal folds. The nerves ramifying the apical epithelium were most dense on the ridges of the mucosal folds. In this layer, the mean density of sensory nerves was 1,900 terminal processes per mm2, and they arose from 709 central fibers per mm2 that had ascended through the epithelium from the basal plexus. The average SP-IR length–density of axons in the apical epithelium was 88 and 22 mm/mm2 in the basal layer. The pig bronchus is more densely innervated than the rat trachea, which has 44 mm/mm2 SP-IR axons in the epithelium, but an apical layer was not distinguished (4). In human bronchi, PGP-IR epithelial nerves were revealed, most strikingly in the 39-year-old nonsmoker, where they were distributed along the length of the bronchus. The arrangement of nerves was very similar to that of the pig. An apical layer of varicose processes terminated in enlarged varicosities, and these fibers encircled goblet cells. These processes arose from fibers that had crossed the epithelium from a basal plexus that was supplied by nerve bundles in the lamina propria. By demonstrating these layers of nerves in the epithelium that are common to the human and pig and by identifying sensory neuropeptides in these porcine nerves, this study has strengthened the case for this arrangement of epithelial sensory nerves in humans. Ultrastructural studies of human bronchial epithelium have also revealed axon profiles situated close to the basement membrane and to the airway lumen, with some fibers laying between the epithelial cells that were characteristic of sensory nerves (35). In our much older patients with a history of smoking, there were fewer PGP-positive epithelial nerves, and they tended to be unevenly spread across the epithelium. SP-IR epithelial nerves were faintly stained in human bronchi. The weak signal suggests low concentrations of SP-IR in the varicosities. This could be due to break down of the tachykinins by neuroendopeptidase, SP being particularly sensitive, during the unavoidable delay before fixation. Some studies have been able to demonstrate a significant number of SP-IR nerves in human bronchial epithelium in thin sections (16, 17, 26). However, in the majority of studies, SP-IR nerves are described as "negative" (25, 27, 28, 36) or "rare" (24, 26) in human bronchial epithelium. It appears that the difference lies mainly in the age and medical history of the subjects (26, 28) and sensitivity of the method. There are obvious limitations in obtaining young airways that have been rapidly fixed. This shows that some improvement in demonstrating SP nerves can be obtained using whole mounts in conjunction with confocal microscopy where high-quantum yield, long-wavelength fluorophores that emit strongly at a wavelength well separated from tissue autofluorescence were employed.
In the pig, SP-IR fibers comprised 94% of all nerves in the epithelium since almost all nerves stained by pan-neuronal marker PGP colocalized with SP. This compares favorably with rat trachea in which SP-IR varicose axons comprised 90% of those staining for PGP (4), and thus, the great majority of nerves were likely to be C-fiber afferents. The residual population of nerves (6–10%) was significant in both studies and may be a subgroup of mechanosensory, thin myelinated A In the pig and human, many of the fibers in the apical epithelium pass between epithelial cells and closely follow around the apex of goblet cells, with varicosities and terminal endings lying on their circumference. This arrangement of apical nerves is likely to facilitate the detection of chemical substances, with these nerves releasing neuropeptides that trigger local protective effects such as increased frequency of cilia beat on epithelial cells and mucous secretion from goblet cells. There is functional evidence that increased beat and secretion is mediated by SP activation of tachykinin receptors (neurokinin 1) present on these cells (46, 47). These local functions are thought to be characteristic properties of bronchial C fibers in addition to evoking centrally mediated reflexes (31, 48). SP-IR and CGRP-IR fibers in the lamina propria were seen passing through capillary beds and to follow arterioles (less than 1 to 10 µm separation); some fibers followed closely the vascular smooth muscle of small arterioles (less than 1 µm). In rat trachea and dog bronchus SP-IR nerve profiles also closely follow arterioles in the lamina propria (4). Thus, neuropeptides released from these fibers can readily diffuse over the vascular smooth muscle to cause vasodilation. These neural structures provide the basis for the functional studies that have shown stimulation of vagal C fibers releases tachykinin neuropeptides in pigs (12) and dogs (13) with an increase in tracheobronchial blood flow; a component of this vasodilation has been attributed to an axon reflex.
We have shown that there is continuity of putative C fibers (CGRP-IR), necessary for an axon reflex, from the epithelium to deep in the lamina propria by tracing DiI-labeled nerves that stained for CGRP-IR. These fibers were followed from their varicose nerve endings in the apical epithelium through the lamina propria directly, or indirectly via large nerve bundles, to vascular beds where they ran along arterioles (less than 1 µm). Identification of some DiI diffused along nerve fibers, tracing their pathways, and also diffused across their membranes into adjacent fibers. Thus, the entire width of nerve bundles and trunks was revealed; these were of similar dimensions as those obtained using PGP-IR. CGRP-IR staining was a minor component of large nerve trunks. DiI also labeled a few cells that were in close contact with the nerves. These cells need to be identified, but antibodies to mast cell tryptase and human leukocyte antigen-DR raised against human antigens were not cross-reactive in pigs. Although the extensive staining of nerves by DiI is useful, an alternative nerve tracer that remains confined to the fiber, which it labels initially in the epithelium, should also be sought for subsequent tracing experiments. SP-IR and DiI/CGRP-IR co-stained fibers were rarely seen close to postcapillary venules. This is consistent with studies in rat trachea and dog bronchus in which SP-IR nerves were not seen near postcapillary venules (4). The paucity of SP nerves near the postcapillary venules could be construed as consistent with the view that leakage from them in large mammals is not a prominent feature of neurogenic inflammation as it is in rodents (51, 52). This is controversial as plasma extravasation can be induced in patients with asthma by SP inhalation (53). Plasma extravasation occurs in the rat via NK-1 receptor-mediated gap formation between endothelial cells (49, 54); the source of SP may be the extensive plexus of C fibers in the basal epithelium (4). In the pig, the postcapillary venules lay 15–20 µm below a plexus of SP-rich nerves in the basal epithelium. We have also shown that human bronchi have a similar neural plexus, but we could demonstrate only traces of SP-IR in these fibers. In summary, the bronchial epithelium of the pig is abundantly supplied by varicose SP-IR and CGRP-IR nerve fibers. These fibers lie in a plexus above the basal membrane and project upward between the epithelial cells to the apical epithelium, where they arborize into processes that terminate in enlarged varicosities adjacent to the airway lumen. The basal layer is supplied by nerve bundles from the lamina propria that arise from nerve trunks that run deep in the lamina propria adjacent to the airway smooth muscle. As these bundles pass through the microvasculature beneath the epithelium, some run along vessels, primarily arterioles, and lie within a micron of their vascular smooth muscle. The continuous nature of these pathways has been demonstrated by tracing individual nerve fibers from the lumen with DiI and provides a structural basis for the existence of a local axon reflex that could be involved in the vasodilation observed in neurogenic inflammation. In human bronchus, a similar distribution of epithelial nerves was shown using the pan-neuronal marker PGP 9.5, but staining for SP-IR was weak or not detected.
The authors thank Dr. Markus Weichselbaum for consultation on confocal imaging and image analysis and to Dr. Alan James, Dr. Neil Carroll, and Dr. Darryl Knight for human lung samples. Jasmine Lamb was the recipient of an Annie Phillip's scholarship.
Supported by an Australian Research Council Grant. Received in original form December 5, 2001; accepted in final form July 30, 2002
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
fibers (rapidly adapting receptors) that are thought to lie in the epithelium (
