© 2003 American Thoracic Society DOI: 10.1165/rcmb.4889 The Three-Dimensional Distribution of Nerves Along the Entire Intrapulmonary Airway Tree of the Adult Rat and the Anatomical Relationship Between Nerves and Neuroepithelial BodiesDepartment of Anatomy, Physiology, and Cell Biology, School of Veterinary Medicine, University of California, Davis, California Address correspondence to: C. G. Plopper, Ph.D., Department of Anatomy, Physiology, and Cell Biology, School of Veterinary Medicine, University of California, Davis, CA 95616. E-mail: cgplopper{at}ucdavis.edu
Using airway microdissection and three-dimensional confocal microscopy techniques in combination with the immunomarkers protein gene product (PGP) 9.5 and calcitonin gene-related peptide (CGRP), we defined the distribution of small afferent nerves fibers and all nerves throughout the intrapulmonary airways, along with the distribution of airway neuroendocrine cells and neuroepithelial bodies. We found (i) the presence of CGRP-and PGP 9.5-positive structures along the entire intrapulmonary airway tree of adult rats, (ii) decreasing nerve density from more proximal to more distal generations of conducting airways, (iii) the presence of nerve fibers in terminal bronchioles, (iv) the asymmetrical distribution of nerves within a single generation of intrapulmonary airway with regard to associated vessels, (v) the frequent interchange of single nerve fibers across epithelial and sub-epithelial compartments without termination, and (vi) a definably intimate relationship between afferent nerves and neuroepithelial bodies (NEBs) (i.e., 58% of NEBs studied were observed to have nerve fibers coursing through them, indicating direct connections). We conclude that the distribution of nervous elements (nerve fibers and neuroendocrine cells) within the intrapulmonary airways is highly heterogeneous, varying between airway levels and locally within a specific airway level.
Abbreviations: calcitonin gene-related peptide, CGRP • fluorescein isothiocyanate, FITC • neuroepithelial body, NEB • neuroendocrine cell, NEC • phosphate-buffered saline, PBS • protein gene product, PGP • standard error, SE
It has long been suspected that small sensory nerves play a central role in a number of airway pathophysiologies, including airway hyper-responsiveness and inflammation. As early as 1967, activated vagal sensory nerve endings were linked to stimuli causing bronchoconstrictive responses in patients with asthma (1). Soon after, the neuropeptides contained within these small sensory nerves were implicated as potentiators of inflammation (2–6). Although it is tempting to draw conclusions regarding the functional significance of individual fibers within the airway, the role of these nerves in the development of airway hyper-reactivity and inflammation cannot be assessed before the development of a firm appreciation for the distribution of these fibers. Furthermore, the anatomical associations between nerve fibers and other airway structures should be explored in detail before speculations functionally linking nerves to other airway cell types, such as neuroendocrine cells (NECs). Pulmonary NECs and pulmonary neuroepithelial bodies (NEBs) are distributed throughout the pulmonary airways of rats and are often found in close proximity to nerve fibers. Pulmonary NECs and NEBs, which are essentially aggregates of pulmonary NECs, contain a number of neuropeptides also produced and contained in small sensory nerves. As early as 1949, Frolich observed NEBs in close association with nerves (7). Lauweryns later confirmed the relationship by electron microscopy for a limited number of NEBs (8). The scattered positioning of NECs throughout the airway, their ability to degranulate in response to changing airway conditions, and their association with nerve fibers suggest that airway nerve/NEC interactions may play an important role in the exaggerated airway reflexes associated with airway diseases (9, 10). Previous work draws attention to the close associations between NEBs (aggregates of NECs) and nerves but fails to define the frequency of such relationships. Weichselbaum and colleagues have created three-dimensional maps of the nerve plexus overlying the smooth muscle on the adventitial surface of intrapulmonary airways in the fetal pig (11). Although their study demonstrated the relationship between nerves and smooth muscle, their approach from the adventitial side of the airway wall hampered the definition of nerve fibers distributed within the epithelial compartment. Baluk and colleagues followed immunoreactive nerve fibers into the epithelial and lamina proprial compartments using two-dimensional techniques. This allowed them to estimate the relative length of nerve axons per surface area of epithelium, lamina propria, or muscle (12). However, it did not enable them to define the continuity of nerve fibers across airway wall compartments. In addition, although their study did recognize epithelial cells that were not nerves but were immunoreactive to protein gene product (PGP) 9.5, a nonspecific neuronal marker that stains NECs, NEBs, and nerves, their study did not identify NEB connections to the local nerve network (12). Shimosegawa and Said, who attempted to define the frequency of NEB interactions with calcitonin gene-related peptide (CGRP)-positive nerve fibers in thick sections, succeeded only in nonpulmonary airways (13). Finally, none of the three followed the distribution of neural structures into the most distal conducting airways or compared differences in local distribution within a single airway generation. Using airway microdissection and three-dimensional confocal microscopy techniques in combination with the immunomarkers PGP 9.5 (a nonspecific neural marker that stains NEBs and nerves) and CGRP (a vesicular stain that labels rat NEBs and small afferent nerves), we attempted to define (i) the abundance and distribution of small afferent fibers and all nerves along the entire adult rat pulmonary airway tree, including terminal bronchioles; (ii) the differences in airway nerve distribution with regard to associated vessels (proximity to pulmonary artery verses pulmonary vein); (iii) nerve fiber continuity within and across the epithelial layer; and (iv) the relationships between nerves and other airway epithelial cell types, specifically neuroepithelial bodies. Our results demonstrate (i) the presence of CGRP- and PGP 9.5-positive structures along the entire pulmonary airway tree of adult rats, (ii) decreasing nerve density from more proximal to more distal generations of conducting airways, (iii) the presence of nerve fibers in terminal bronchioles, (iv) the asymmetrical distribution of nerves within a single generation of pulmonary airway, (v) the frequent interchange of single nerve fibers across epithelial and sub-epithelial compartments without termination, and (vi) a definably intimate relationship between nerves and NEBs, of which there was a clear direct connection to 58% of the NEBs studied.
Male Wistar rats (300–350 g) were obtained from Charles River Breeding Laboratories (Wilmington, MA). All animals were housed in animal facilities for at least 5 d before use and were provided food and water ad libitum. Animals were anesthetized with sodium pentobarbital (200 mg/kg), tracheotomized, and killed by exsanguination. The lungs were fixed in situ via infusion through tracheal cannula with 2% paraformaldehyde at 30 cm of pressure for 90 min. After fixation, the right middle (cardiac) lobe was removed to phosphate-buffered saline (PBS) and microdissected to expose the axial path of the intrapulmonary airway tree. Costal and mediastinal halves were preserved for immunohistochemistry.
After fixed tissue had been microdissected and excess parenchyma trimmed, the airway halves were washed three times in dimethyl sulfoxide (10 min per wash). After a 10-min wash in PBS (pH 7.2), the tissue was immersed in solutions of primary antibody (dilutions: rabbit
The antibodies to PGP 9.5 and CGRP were obtained from Biogenesis (Poole, UK). The secondary antibodies (anti-rabbit and anti-goat) were conjugated to FITC, tetramethyl rhodamine isothiocyanate, or indocarbocyanine (Jackson Laboratories, West Grove, PA). Fluorochrome-labeled immunoglobulins had been pre-absorbed to minimize nonspecific reactions with rat tissues or serum proteins from the host of the unintended primary antibody. Negative controls were performed by selectively replacing primary antibodies with PBS and incubating as before with secondary antibodies.
Epifluorescent images of the nerves, NEBs, and NECs labeled with FITC, rhodamine isothiocyanate, or indocarbocyanine were obtained using 4x to 60x water immersion objectives (Olympus W Plan; Melville, NY) and confocal laser scanning microscopes (MRC 600 and 1024; Bio-Rad, Hercules, CA) with COMOS software (version 3.0 and 3.2; Bio-Rad). The airway whole mounts were optically sectioned by scanning at increasing depths of focus (steps varied between 1 µm and 20 µm depending on the water immersion objectives used). Aperture settings were chosen to minimize overlap between consecutive optical sections. Image processing was done with Adobe Photoshop 5.5 software.
In eight male Wistar rats, the ratio of PGP 9.5- and CGRP-positive nerves to luminal epithelium were estimated using an unbiased optical "dissector" in an optical series of confocal images captured using a 40x objective at specified airway locations (14). The surface area of nerves and luminal epithelium were estimated by applying a quadratic lattice to each section and counting the number intersections with each object of interest in x and y directions. Because the points of the lattice are rays in the z direction sweeping through space, intersections in the z direction were recorded by transitions within and without the epithelial or nerve surface between confocal images. Objects that intersected the top section and two sides of the counting frame were excluded. Stacks of serial sections varied from 14–50 images, depending on the orientation of the airway surface in the dissected whole mounts. For statistical comparison between airway generations comprising the length of the airway, only images from the mediastinal side were analyzed. The generation number was determined by direct count of each branch on both halves of the dissections.
We analyzed the surface area of PGP 9.5-positive nerves per surface area of luminal epithelium and the surface area of CGRP-positive nerves per surface area of luminal epithelium using analysis of variance, where airway generation number was the grouping factor. Post hoc analysis between the seven groups were completed using Fisher's least significant difference test (Statview, Version 5.01; SAS Institute, Cary, NC). We compared the density of PGP 9.5-positive nerves on opposing airway surfaces (mediastinal versus costal) using the Student's t test (Statview, Version 5.01; SAS Institute). All data are presented as mean ± standard error (SE). Statistical significance was considered at P < 0.05.
To define the density and distribution of nerve fibers and NEBs within pulmonary airways, whole-mount preparations were labeled with antibodies to PGP 9.5, a nonspecific neuronal marker of ubiquitin carboxyl terminal hydrolase (15–18), and CGRP, a neuropeptide product of small afferent nerves and neuroendocrine cells (19, 20). Figure 1 demonstrates nerves, a neuroendocrine cell, and a NEB double labeled for PGP 9.5 and CGRP. PGP 9.5 labels neural structures uniformly, whereas CGRP labels vesicles within nerves and focal areas of intensity within the neuroendocrine cells and NEBs. The fluorochromes attached render the PGP 9.5 signal green and the CGRP signal red.
In all fields observed, nearly all nerves positive for PGP 9.5 contained a nerve fiber positive for CGRP. Figure 2 compares PGP 9.5- and CGRP-positive structures within the airway wall and identifies segments of PGP 9.5-positive nerves that do not label for CGRP. Because of the punctate nature of the CGRP signal compared with the PGP9.5 signal, meaningful morphometric comparisons between the density of CGRP-positive nerves and PGP 9.5-positive nerves were not possible.
The estimated density of airway nerves varied significantly from airway generation to airway generation. The densest nerve plexuses were observed in more proximal airways, with progressively finer and looser plexuses observed in more distal airways (Figure 3). When stacks of confocal images taken with a 40x lens were analyzed in eight animals at a variety of airway generations spanning the entire intrapulmonary airway tree, the mean density of PGP 9.5-positive nerves varied significantly between airway generations. There was a significant decrease in the density of nerves between the third generation and the fifth generation (mean difference 0.251, P = 0.0454), between the fifth generation and the seventh generation (mean difference 0.272, P = 0.0305), and between the seventh generation and the most distal airways (mean difference 0.522, P = 0.0003). The density of nerves located in the terminal bronchioles was approximately one fourth that of the nerves found in the most proximal intrapulmonary airway generations (Figure 4). Simple regression analysis of individual animals consistently showed an inverse relationship between airway generation number and the density of PGP 9.5-positive nerve fibers (mean R value 0.921, SE = 0.021; graphs not shown). The density of CGRP-positive nerves also varied significantly between a number of airway generations. We measured a significant decrease in the mean density of nerves between the fifth and ninth airway generations (mean difference 0.228, P = 0.0338) and between the ninth generation and terminal bronchioles (mean difference 0.261, P = 0.0251) (Figure 5). Simple regression analysis of individual animals consistently showed an inverse relationship between airway generation number and the density of CGRP-positive nerve fibers (mean R value 0.862, SE 0.045; graphs not shown). Valid comparisons between densities of nonspecific PGP 9.5-positive nerves and small afferent (CGRP-positive) nerves were not possible because of the nature of our markers (vesicular verses whole cell).
Our technique allowed us to define asymmetries in density between mediastinal and costal sides of the same airway. By documenting the airway position, we could evaluate nerve density on artery versus the artery-deficient airway halves. In the rat, the pulmonary artery and bronchus follow a parallel course (21). The mediastinal half displayed a denser distribution of PGP 9.5-positive nerve fibers (Figure 6). The mean density of PGP 9.5-positive nerves (n = 8) on opposing walls within the third airway generation varied significantly (mean difference 0.608, P = 0.0011). No significant differences in the distribution of CGRP-positive fibers on opposing airway halves were identified.
To better appreciate the continuity of nerves between the epithelial and interstitial compartments, we used optical sections to infer the location of nerves relative to NEBs and NECs. Because NECs, or a portion of the NECs that make up NEBs, are located on or near the basement membrane, we used their position to estimate where nerves cross from the interstitial to the epithelial compartments based on planar reference to NECs. The projected series of images in Figure 7 represents a 20-µm section of the bronchiolar wall beginning within the epithelial compartment and extending into the interstitium. A number of nerve segments visible within the epithelial compartment abut nerve segments, which lie within the submucosal compartment. Figure 7 demonstrates the consistent movement of nerve fibers between epithelial and interstitial planes. Final termination of a nerve fiber within the epithelium was observed in less than  5% of the nerves followed.
All NEBs were observed in close proximity to a nerve fiber; however, nerves coursed into or through NEBs in only 15 of the 26 NEBs analyzed in detail (58%). Nerve fibers rarely terminated in close proximity to NEBs. The NEB represented in Figure 1 was positive for CGRP and PGP 9.5, as were all NEBs observed. In fifteen cases, we were able to follow the passage of one or more nerve fibers through a NEB by scanning through the structure (Figure 8A). We interpret nerves that course through NEBs as being directly connected to those NEBs. In the remaining 11 cases, we observed nerves in association with (but not in direct contact with) the NEBs (Figure 8B).
Most NEBs were identified in close association with the more prominent fibers of their associated airway generation. For this reason, NEBs were often sequentially recorded along the airway in alignment with thicker central nerves (Figure 9). Nerve fibers did not seem to approach clusters of NEB in accordance with an established pattern, however. Nerves approach NEBs from more proximal and more distal portions of the airway.
PGP 9.5 and CGRP colocalized to many nerves found in the walls of the conducting airways, indicating an abundance of afferent fibers within the intrapulmonary airway tree. As reported previously, antibody to PGP 9.5, a nonspecific neural marker for ubiquitin carboxyl terminal hydrolase, uniformly labeled nerve fibers and ganglia (17). Antibody to CGRP, a sensory neuropeptide, had a varicosed labeling pattern indicative of CGRP packaging in synaptic vesicles. PGP 9.5 and CGRP antibody labeled NECs and NEBs. PGP 9.5 antibody evenly labeled NECs, the primary component of NEBs, and CGRP antibody exhibited focal areas of intensity within the cells that compose neuroepithelial bodies. Differences in the distribution of the nonspecific neural marker PGP 9.5 and our indicator of small afferent nerve fibers (CGRP) can be seen in NEBs, NECs, and nerve fibers by examining Figures 1A, 1B, and 1C. Using three-dimensional technology to image along the length of the entire intrapulmonary airway tree, we observed a reduction in the density of nerve fibers as we progressed from proximal to more distal airway positions. At selected airway locations, we captured stacks of high-magnification images and used these to estimate the surface area of nerves per surface area of luminal epithelium. Previous research has quantified an inverse relationship between airway nerve density and airway position (with regard to distance from proximal trachea). In proximal airway locations, including the rostral trachea, caudal trachea, and main stem bronchus, a significantly lower nerve density was calculated from sections of more distal nonpulmonary airway positions (12, 13). Our study extends past airway nerve analysis studies and shows dramatic changes in nerve density/positional relationships within intrapulmonary airways. We examined opposing surfaces of an airway within a single generation and observed distinct differences in nerve density. The mediastinal half the rat airways had a 72% increase in nerve fiber density and thicker central fibers compared with the costal airway half. These differences in nerve distribution could easily contribute to differences in the biology of airways within the same airway generation. Any function dependent on innervation or neurochemical modulation would be affected by this heterogeneity in nerve distribution to include mucus secretion, glandular function, smooth muscle tone, vascular permeability, inflammatory recruitment, and epithelial repair (2, 3, 22, 23). Furthermore, in diseased airways, a state usually associated with imbalances, one might expect additional divergence in nerve distribution. For this reason it is critical that investigators and clinicians carefully define their site of sample or biopsy collection, especially when the mechanism of the disease under investigation might involve the influence of local nerve networks. Our results lead to the question of whether other cells types, including mucous cells, vascular plexuses, glands, and inflammatory cells, are influenced by the asymmetrical distribution of nerve fibers. Our study emphasizes the continuity of the nerve network between epithelial and submucosal layers throughout the intrapulmonary airway tree. Individual nerve fibers frequently enter the epithelial compartment and then return to the submucosal compartment. Based on the early discovery of intra-epithelial nerve endings, investigators have predicted the termination of submucosal nerve fibers within the epithelial layer or within the lumen of the airway (24). Although bulb-like nerve endings were present within the epithelium, more frequently epithelial nerve fibers traversed the basement membrane, returning to the submucosal layer without termination. Often, a single nerve fiber crossed the basement membrane several times as it moved between epithelial and submucosal layers. It is possible that the varicosed nature of many epithelial nerves may have given the impression of nerve termination within the epithelium in two-dimensional images. The infrequency of nerve fiber termination within the epithelium does not discount their sensory role within the epithelium. Receptors have been clearly identified on unmyelinated sensory axons (25, 26). The continuity of the airway nerve network among airway levels makes it the perfect intermediary for communication between a number of cell types and tissue systems. In the rat, PGP 9.5 and CGRP are present in NEBs and individual NECs of the intrapulmonary airway. The distribution of NEBs in the rat has been mapped previously (9, 13). We emphasized the relationship between NEBs and nerves first described by Frolich using light microscopy (7) and later by Lauweryns using electron microscopy (8). The close proximity of nerves to NEBs has led a number of investigators to suggest close communication between the two, but a definitive explanation of the nature of the interaction in not available. Because NEBs morphologically resemble carotid bodies and because they release vesicular products in response to hypoxia, a sensory (chemoreceptor) function has been postulated for NEBs (8, 27, 28). If this were the case, then one would expect that all NEBs would be innervated by sensory fibers and that degranulation of NECs within NEBs would produce a measurable systemic response. Ultrastructure studies in rabbits indicate that only one third of NEBs are innervated, and denervation studies suggest that only two thirds of those NEBs are innervated by nerves whose cell bodies are housed in the vagal ganglia (29). Assuming that the innervating vagal afferents synapse within the nucleus tractus solitarius (like carotid body vagal afferents), then NEB degranulation would be associated with adjustments in respiratory center function to maximize the partial pressure of oxygen within the airways. Current studies do not provide evidence of a systemic response to NEB degranulation. There does seem to be a local compensatory response to chronic hypoxia, NEB hyperplasia, which is evident in children with hypoventilation syndrome (30). It is unclear whether nerves, afferent or efferent, play a role in this airway NEB hyperplasia. Although past studies predict only a 22% innervation rate of intrapulmonary NEBs by vagal sensory nerve fibers (29), NEB analysis in denervation studies do suggest close interactions between NEBs and afferent nerves. Studies by Lauweryns indicate that NEB activity may be modulated by local reflexes within the airway. When airways containing biogenic amine-positive NEBs are unilaterally deprived of their afferent and efferent innervation via vagal ligation below the nodose ganglia, NEBs become less responsive to airway hypoxia within 3 d, which is sufficient time to deplete axon terminals of neurotransmitter. However, the response of NEBs to hypoxia is not affected when the vagus is ligated above the nodose ganglia (10). These studies suggest that the release of neuropeptides from afferent nerves whose cell bodies are located in the nodose ganglia modulate NEB degranulation in response to hypoxia. Based on detailed three-dimensional analysis of 26 NEBs in male adult rats, we found a definable direct connection between NEBs and nerve fibers in slightly over one half of the NEBs studied. Most of these nerves were positive for CGRP and PGP 9.5. Earlier work by Van Lommel using anterograde neural tracer injected into the nodose ganglia established the existence of NEB/afferent nerve fiber complexes (31). However, the frequency of such relationships was not determined. In neonatal dogs, several NEBs have been found to be deficient of any kind of anatomic innervation (32). Moreover, the aberrant nerve morphology, possibly indicative of degeneration, associated with 40% of the NEB/nerve complexes studied led the investigators to propose that NEB-associated nerves are lost with age. Determining whether rats lose NEB innervation with age or whether such innervation never existed in 40% of the pulmonary airway NEBs was not the purpose of our study. However, we do believe that the paucity of NEBs with definable innervation in adult rats brings into question the contribution of NEB/nerve interactions in normal respiratory function. Conversely, although formidable distances between NEBs and nerves may make rapid communication more difficult, these distances do not necessarily discount interactions between NEBs and nerve fibers. Studies by Jan and Jan demonstrate that peptides released from nerve terminals can influence neurons hundreds of micrometers away (33). Although we may not fully comprehend the interactions of NEC/NEBs and nerves, we know that the availability of neuropeptides at any loci within the airway is for the most part dependent upon these cells. Changing neuropeptide availability can influence local airway biology, possibly enhancing epithelial cell proliferation (23) or exacerbating defense mechanisms by increased blood flow, mucous secretion, airway smooth muscle tone, and/or the recruitment and activation of inflammatory cells (2, 3). The increasing density of neuropeptide-containing structures (like NECs and NEBs) in a number of pathologies hint of their importance. In infants diagnosed with sudden infant death syndrome (34), adults with bronchial asthma (35), and adults with chronic bronchitis (36), notable increases in airway NECs and NEBs were recorded. Because the innervation of these structures was not considered, we cannot predict their local influence. The activity and growth of NEBs may be normally inhibited by innervating nerves. A loss of innervation might contribute to the uncontrolled activity and growth of neuroendocrine cells. The speculation over regional neuropeptide influence in pathology cannot fully be appreciated until the frequency of direct connections between nerves and NEBs are compared among a number of respiratory pathologies. The ratio of nerves to NECs and NEBs may be more significant in the development of airway pathology rather than either marker independently.
This work was supported by NIEHS grants 06791, ES05707, ES00628, ES06700, and T32 HL07013. The authors acknowledge the help and advice of Dr. R. Paige and A. Weir. Received in original form April 26, 2002 Received in final form November 12, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
PGP 9.5, 1/100 and/or goat 
