INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Using retrograde tracing techniques, the predominant afferent supply to the guinea pig airway has been shown to be provided by the vagus nerve (1). The cell bodies of the vagal afferent fibers are located in either of two embryologically distinct ganglia, namely the jugular and nodose ganglia. Combining retrograde tracing with immunohistochemistry, approximately 30% of the afferent fibers innervating the guinea pig trachea have been found to contain tachykinins, substance P (SP), and neurokinin A (2). Virtually all of these neurons are small-diameter neurons located mostly in the jugular ganglia. Only about 1% of the cell bodies of airway afferent fibers located in the nodose ganglia contain SP. Another distinction between nodose and jugular neurons innervating the guinea pig trachea is that the nerve fibers located in the epithelium are almost exclusively derived from cell bodies situated in the jugular ganglia (3).

Physiological studies have revealed that nerve endings arising from the jugular ganglia are analogous to the so-called nociceptive fiber type, where they are only moderately sensitive to mechanical force, but respond vigorously to capsaicin, bradykinin, hydrogen ions, and changes in osmolarity (2, 4, 5). The airway nerve endings arising from the nodose ganglia, by contrast, do not respond to these inflammatory-associated stimuli, but rather are exquisitely sensitive to mechanical force (2). Considering the anatomical data with the physiological studies, it appears that SP-containing fibers in guinea pig airways are nociceptive-like C fibers that arise selectively from the jugular ganglia and project to the epithelial layer.

Many studies have demonstrated quantitative increases in the tachykinin content of afferent nerves innervating visceral tissues caused by inflammatory processes (6). This can also be observed in the airways, where allergic inflammation leads to substantial increases in the content of sensory neuropeptides (9). It has recently been reported in the somatosensory system that inflammation not only causes quantitative changes in sensory neuropeptides, it is also associated with a qualitative change in tachykinin neurobiology, such that SP production is induced in the fast conducting A-fiber type (10). Nerve growth factor (NGF) is known to be elevated in allergically inflamed airways (11, 12). Moreover, NGF is capable of regulating the expression of tachykinin genes in adult sensory neurons (13). Therefore, in the present study, experiments were designed to address two hypotheses. First, that NGF in the airways could lead to increases in the expression of tachykinins in airway-specific sensory neurons and second, that NGF could lead to a qualitative change in the nature of the sensory neurons expressing tachykinins.

The results support the hypothesis that NGF may serve as a signaling molecule in the transduction of inflammation by elevating sensory neuropeptide production in the airways. We also provide evidence that there is a phenotypic change in the nature of the tachykinergic neurons carried by the vagus nerve after NGF exposure, such that SP production is induced in large-diameter capsaicin-insensitive nodose neurons innervating the airways.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Male guinea pigs (150 to 250 g; Harlan Sprague-Dawley, Inc., Indianapolis, IN) were used to study the effect of NGF on tachykinin expression of airway-specific sensory neurons. On Day 1, 30 µl of a retrograde tracer (fast blue) was instilled into the upper trachea of the guinea pig. After 7 d, the trachea of control and experimental animals were injected with 50 µl of vehicle (normal saline) or NGF- (1 µg/µl), respectively. After 24 h, control and NGF-treated guinea pigs were killed (overdose of carbon dioxide) and the nodose and jugular ganglia were removed and prepared for SP- and neurofilament (NF)-immunohistochemistry. Thus, the experimental design consisted of two groups: a control group of four guinea pigs that received instillation of fast blue and 50-µl injection of saline into the proximal trachea and an experimental group of four guinea pigs that received 30 µl instillation of fast-blue and 50-µl injection of NGF- into the proximal trachea.

Tracheal Instillation of Retrograde Tracers and NGF Injection

Guinea pigs were anesthetized by an intramuscular injection of ketamine hydrochloride (50 mg/kg) and xylazine hydrochloride (2.5 mg/kg). A volume of 30 µl of 5% fast blue (Sigma Chemical Co., St. Louis, MO) was instilled into the proximal region of the trachea using a Hamilton microsyringe with a curved metal feeding tube connected to the syringe. In addition, flexible wire covered with gauze was connected to a plexiglass board elevated 45°. The guinea pig was placed on the board and the gauze wire was positioned behind the upper incisors to aid in the extension of the head and the opening of the mouth. Gauze-covered forceps were used to deflect the tongue outward and ventral. A light was positioned for optimal visualization of the posterior oral cavity. The feeding tube connected to the microsyringe was deflected downward past the posterior region of the tongue and into the laryngeal orifice. Once an increase in resistance was felt, the feeding tube was pushed through the laryngeal orifice and the tracer was immediately instilled into the proximal trachea. Upon entrance into the trachea, the microsyringe was rotated in a circular pattern to enable a greater chance of tracer distribution around the circumference of the trachea. The animal was then removed from the plexiglass board and positioned supine during recovery from anesthesia. With this technique, we have previously noted that the fluorescent dye is limited to the rostral region of the trachea (3). Seven days later the animal was again anesthetized, and using sterile techniques, a small incision was made in the rostral portion of the trachea. Using a Hamilton syringe, 50 µl of NGF- (1 µg/ml), or the diluent (normal saline), was injected circumferentially into the tracheal wall.

Tissue Removal and Preparation

One day after saline or NGF injection, the experimental and control animals were killed by 100% CO₂ asphyxiation and exsanguinated. The right nodose ganglion and right jugular ganglion were removed. All tissues were fixed in picric acid-formaldehyde fixative for 3 h at 4° C. The fixative consisted of 2% paraformaldehyde, 15% saturated picric acid, and 0.15 M phosphate buffer. The tissues were rinsed 3 times with a 0.1 M phosphate-buffered saline (PBS) containing 0.3% Triton-X-100 (PBS-Tx). During the final rinse, the tissues remained in the PBS-Tx overnight at 4° C.

The nodose and jugular ganglia were mounted with a specific orientation on corks so that the first section collected was of the ventral surface. The ganglia were frozen in isopentane, cooled with liquid nitrogen, and stored in airtight plastic bags at 60° C.

Continuous serial cryostat sections (12-µm thickness) of the nodose and jugular ganglia were collected sequentially on four different subbed coverslips. The first coverslip had sections 1, 5, 9, . . . , the second coverslip had sections 2, 6, 10, . . . and so on. The entire ganglia was sectioned, however not every section was evaluated. The retrograde labeling by fast blue was analyzed on all evaluated sections. Coverslips 1 and 2 were used for SP immunocytochemistry. Coverslips 3 and 4 were used for NF and SP colocalization immunohistochemistry studies and cell diameter determination of the colocalized neurons.

SP Immunocytochemical Procedures

Immunocytochemical procedures for localizing a single antigen are identical to those described previously (3). Briefly, cryostat sections on gelatin-coated coverslips were covered with a diluted primary antiserum, rabbit anti-SP (Peninsula Laboratories, Belmont, CA; 1:200 working dilution). The coverslips were incubated in a humidified chamber at 37° C for 30 min. The coverslips were rinsed with a 1% bovine serum albumin-phosphate saline buffer containing triton-X solution (PBS-Tx + BSA) three times, allowing 5 min for each rinse. The sections on the coverslips were covered with a diluted secondary antiserum, fluorescein isothiocyanate-labeled goat anti-rabbit (Peninsula Laboratories, 1:200 working dilution) and incubated at 37° C for 30 min. The coverslips were rinsed again in PBS-Tx + BSA, three times for 5 min. The coverslips containing alternate sections of nodose or jugular ganglia were mounted using fluoromount. Controls for specificity of primary antisera consisted of absorption of 1 µg/ml antisera with SP. Nonspecific background labeling was determined by omission of primary antiserum. Even after these controls were employed, however, cross-reactivity of antisera with other known or unknown peptides present in the tissue cannot be excluded by immunocytochemical procedures. Therefore, the term peptide-immunoreactivity (i.e., SP-IR) is used here when referring to labeled structures.

Without further processing, the sections were analyzed for the distribution of fast blue-labeled vagal ganglia neurons containing SP-IR. Twenty-four sections were observed and used to count the total number of fast blue-labeled nerve cell bodies containing SP-IR and the overall number of nerve cell bodies in each vagal ganglia labeled with fast-blue. The sections were counted by direct observation using a laboratory cell counter and only counted neurons with a defined nucleus. The percentage of fast blue-labeled neurons containing SP-IR was calculated for each vagal ganglion by dividing the number of fast blue- labeled cell bodies containing SP-IR by the total number of fast blue- labeled cell bodies.

SP and NF Colocalization Immunocytochemical Procedures

Previous studies showed that all NF-positive neurons were SP-negative and essentially all NF-negative neurons were SP-positive (2). Therefore, we used NF-IR as a marker for large, myelinated A fibers and SP-IR as an indicator for small, unmyelinated fibers by colocalizing these two antigens (2). The immunofluorescence for SP was described previously. The primary antisera used for NF-IR (160 kD) was a mouse monoclonal anti-NF (Boehringer Mannheim, Indianapolis, IN; Clone NN18, diluted 1:30). The mouse antibody was labeled by a biotinylated goat-anti-mouse IgG (Amersham Life Science, Piscataway, NJ; diluted 1:50) followed by a streptavidin-Texas Red conjugate (Molecular Probes, Eugene, OR; diluted 1:100).

These sections were analyzed for the distribution of fast blue- labeled vagal (nodose or jugular) ganglia neurons containing colocalized SP- and NF-immunoreactivity. A sample population of 200 fast blue-labeled neurons representing four ganglia (nodose or jugular) from saline- and NGF-treated animals were randomly selected for observation. The percentage of fast blue-labeled vagal neurons containing colocalized SP-IR and NF-IR were calculated by dividing the number of fast blue-labeled cell bodies, nodose or jugular, containing SP-IR and NF-IR by the total number of fast blue-labeled cell bodies, nodose or jugular. The same criterion was followed to identify and count neurons as described previously.

In addition to identifying the colocalized vagal labeled neurons, the neuronal diameter was also measured from this population of vagal neurons. The long axis of the nucleated cell body and the short axis perpendicular to it were measured by using a calibrated ocular grid. The mean of the long and short axis was taken as the "neuronal diameter" throughout processing the data.

Determination of Sensory Fiber Conduction Velocities

The conduction velocities of nodose and jugular nerve fibers with mechanosensitive receptive fields in the trachea were determined as previously described (2). Briefly, the trachea was isolated with the right vagus nerves and associated sensory ganglia intact from saline-treated guinea pigs. The trachea and vagal ganglia were separately superfused with Krebs bicarbonate buffer solution (maintained at 37° C) and gassed with a combination of 95% oxygen and 5% carbon dioxide. Single fiber activity was discriminated by placing a concentric electrical stimulating electrode on the recurrent laryngeal nerve, through which the majority of fibers enter the trachea. The recording electrode was placed within the ganglion and manipulated until a single unit activity was detected. When electrically evoked action potentials were seen, the stimulator was switched off and the trachea and bronchi were gently brushed with a von Frey filament. Mechanically sensitive receptive fields were revealed when a burst of action potentials was elicited in response to von Frey filament stimulation. Conduction velocities were calculated by electrically stimulating the receptive field with a single 0.1-ms square pulse of suprathreshold current using a concentric ring electrode, and then measuring the distance traveled along the nerve pathway divided by the time between the shock artifact and the recorded action potential.

Data Analysis

The counting of all neurons was carried out by two investigators in an unbiased manner. The counting was carried out in a "blind" fashion, i.e., the investigator counting the neurons did not know the treatment status of the ganglion studied. The differences among the treatment groups in the percentage of cells that contained SP-immunoreactivity were analyzed for statistically significant differences using a chi-squared analysis (Statview II program; Abacus Concepts, Lafayette, CA).

Photomicrographic Illustrations

After the coverslips of the nodose and jugular ganglia were mounted with fluoromount, the sections were observed using an Olympus DX60 fluorescence microscope equipped with fluorescein (excitation 455 to 500 and emission 515 to 535 nm), rhodamine (excitation 520 to 550 nm and emission 590 to 620 nm), and ultraviolet (excitation 330 to 385 nm and emission 420 to 460 nm) filters. The sections were photographed with an Olympus PM-C35PX camera. The slides were scanned into Adobe PhotoDeluxe 1.0 program for the Power Macintosh using the Polaroid SprintScan 35 system. Digital images were printed on an Epson 700 Stylus Printer without additional image processing.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nodose Ganglion Neurons Innervating the Trachea

In nodose ganglia obtained from control (normal saline-treated) guinea pigs, 838 neurons were retrogradely labeled from the trachea. Among these fast blue-labeled neurons only 2.9 ± 0.53% (24 of 838) were SP-IR. Although less than 3%, this value is larger than the 1.0 ± 0.05% (4/399) and 1% (3/267) we obtained in two of our previous studies (2, 3). This may be due to the fact that in the present study the animals were subjected to a minor surgical procedure for normal saline injection. In nodose ganglia obtained from guinea pigs in which the trachea was treated with NGF- (1 µg/µl), a significant 3.6-fold increase was noted in the percentage of neurons that were SP-IR, such that now 10 ± 1.07% (52 of 492) of the fast blue- labeled neurons were SP-positive (Figure 1).

View larger version (27K): [in this window] [in a new window]   Figure 1.   The percentage of fast blue-labeled nodose neurons containing SP-immunoreactivity 24 h after saline or NGF treatment. An asterisk denotes a significant change from saline- to NGF-treated guinea pigs (p  0.05). Each value is a mean of four animals ± SEM.

We have previously reported that tracheal-specific neurons in the nodose ganglia that were SP-negative, were nearly uniformly large in diameter, and stained positively with antibodies against NF-160 kDa protein (2). We therefore addressed the hypothesis that NGF caused a phenotypic switch in the tachykinergic innervation such that large neurons that are NF-IR also become SP-positive (Table 1). We randomly sampled sections from four ganglia obtained from saline-treated animals, analyzing the first 100 nodose neurons that were retrogradely labeled from the trachea. As expected, of these 100 neurons, 98 were NF-IR and had relatively large (> 30 µm) somal diameters. Consistent with our previous findings, none of the 98 NF-positive neurons were SP-positive (Figure 3). Two of 100 nodose neurons that stained positively for SP were NF-IR negative, and had small somal diameters (18 and 20 µm), indicative of C-fiber neurons. In nodose ganglia obtained from animals that received tracheal injection of NGF, 96 of 100 retrogradely labeled nodose neurons had large somal diameters and were NF-IR. Among this large-diameter NF-positive population, eight were also clearly positive for SP-immunoreactivity (Figure 4). This is consistent with the induction of SP production in approximately 10% of the large-diameter airway-specific nodose neuron population.

View larger version (70K): [in this window] [in a new window]   Figure 3.   Photomicrographs of a nodose cell body (arrow) labeled by retrograde tracing of fast blue from the tracheal wall (A) showing the localization of NF-positive immunoreactivity (B) and SP-negative immunoreactivity (C ) after 24 h saline (control) treatment. Scale bar = 20 µm.

View larger version (36K): [in this window] [in a new window]   Figure 4.   Photomicrographs of a nodose cell body (arrows) labeled by retrograde tracing of fast blue from the tracheal wall (A) showing the localization of NF-positive immunoreactivity (B) and SP-positive immunoreactivity (C ) after 24 h NGF treatment. Scale bar = 20 µm.

Conduction velocity. We measured the conduction velocity of 264 nodose fibers with mechanically sensitive nerve endings in the trachea. Only 10 fibers had conduction velocities in the C-fiber range of  1 m/s. The conduction velocity of these 10 fibers averaged 0.7 ± 0.1 m/s. The remaining 254 fibers had conduction velocities in the A range that averaged 5.0 ± 0.2 m/s.

Jugular Ganglia Neurons Innervating the Trachea

We investigated a total of 867 airway-specific (retrogradely labeled) neurons in the jugular ganglia. Consistent with our previous reports (1, 2), approximately 40 ± 1.0% of the jugular neurons projecting fibers to the trachea were SP-positive. This increased slightly to 56 ± 3.1% in NGF-treated animals (Figure 2). To address the hypothesis that NGF induced a phenotypic switch in the type of SP-positive neurons of the jugular ganglion, we evaluated the cell diameter and NF-immunoreactivity.

View larger version (34K): [in this window] [in a new window]   Figure 2.   The percentage of fast blue-labeled jugular neurons containing SP-immunoreactivity 24 h after saline or NGF treatment. An asterisk denotes a significant change from saline- to NGF-treated guinea pigs (p  0.05). Each value is a mean of four animals ± SEM.

In saline-treated animals 47 of 100 jugular neurons retrogradely labeled from the airway had relatively large somal diameters (> 30 µm), and were NF-IR (Table 1). None of these 47 NF-positive neurons were positive for SP-immunoreactivity. Among the 53 NF-negative neurons, by contrast, all were SP-positive, and had relatively small diameters, averaging 20 ± 2 µm. In jugular ganglia obtained from animals in which the trachea was previously injected with NGF, 39 of 100 neurons had large somal diameters and were NF-positive. Among these neurons, 13 were also SP-positive. This NGF-induced population of SP-positive neurons had cell diameters that were significantly larger than the diameters of the SP-positive population of neurons in the saline-treated animals (33 ± 2 µm versus 20 ± 2 µm, Table 1).

Conduction velocity. The conduction velocities of 231 jugular fibers with mechanically sensitive receptive fields in the trachea/bronchus were investigated. Among this population, 124 fibers (53%) conducted action potentials < 1 m/s with an average velocity of 0.8 ± 0.02 m/s. The remaining fibers conducted action potentials in the A range (5.6 ± 0.3 m/s).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we have demonstrated that NGF, a molecule elevated during airway inflammation (11, 12), not only increases the percentage of airway-specific sensory neurons that express SP-IR, but also leads to a phenotypic switch in the nature of vagal sensory neurons producing this sensory neuropeptide.

Retrograde labeling studies have demonstrated that the vast majority of sensory nerves innervating the trachea and central airways in mammals travel in the vagus nerves. In guinea pigs, approximately half of these fibers emanate from cell bodies located in the nodose ganglia and the other half from cell bodies located in the jugular ganglia (1). Studies on the development of the sensory nervous system have revealed that the nodose and jugular ganglia are distinct structures from an embryological viewpoint, with the former having a placodal origin and the latter arising from the neural crest (14, 15). A growing body of evidence is accumulating to reveal that nerve fibers from these ganglia are also anatomically and physiologically distinct. Consistent with previous studies in the guinea pig (1) and rat (16), we found that SP neurons innervating the guinea pig airways were derived nearly exclusively from the jugular ganglia. In normal saline-treated animals, approximately 50% of the jugular neurons projecting to the trachea were SP-IR. This correlates well with our observation that about 50% of the jugular fibers projecting to the trachea conducted action potentials in the C fiber range. Also supporting the contention that SP-IR neurons project C-fibers to the airways was the observation that SP-IR was found in small neurons with diameters of only about 20 µm that were not immunoreactive to NF. We have previously reported that the jugular C fibers projecting to the airways are only modestly sensitive to mechanical stimulus, but are strongly activated by stimulus such as bradykinin, capsaicin, decreases in pH, and increases in osmolarity (2, 4, 17). These data support the hypothesis that under normal conditions the tachykinergic innervation of guinea pig trachea is derived from jugular C fibers that are analogous to the so-called nociceptive fibers in the somato sensory system.

In the NGF-treated airways, SP production was induced in a population of airway-specific nodose ganglion neurons. These SP-IR nodose neurons had somal diameters that were significantly larger than the SP-IR neurons observed in saline-treated animals. These SP-positive nodose neurons have relatively large somal diameters, which is also consistent with the hypothesis that their fibers conduct action potentials at velocities in the A range. Supporting this contention is our finding regarding the conduction velocity of the 264 nodose nerve fibers studied with receptive fields in the trachea; 97% of them conducted action potentials in the A range. Elegant studies in the dorsal root ganglia have correlated immunoreactivity to NF protein with somal diameter and conduction velocity (18, 19). Thus, neurons projecting C fibers were typically poorly stained with antibodies to NF, whereas large neurons with fast conducting fibers stained intensely for NF protein. We have made similar correlations in the guinea pig vagal sensory ganglia (2). Therefore, also supporting our hypothesis that NGF induces SP production in fast conducting neurons is the observation that the percentage of NF-positive neurons that produce SP increase after NGF exposure from 0% to nearly 10%. It should be noted that fast-blue dye was instilled into the tracheal lumen, whereas NGF was injected into a very localized region of the trachea wall and thus a portion of the fast blue- labeled nerve population may not have been exposed to effective concentrations of NGF. If this is true, 10% may underestimate the efficacy of NGF at inducing SP production in nodose neurons. In any event, the present results lead logically to the hypothesis that after NGF exposure, there is a phenotypic change in airway-specific neurons such that large-diameter neurons that project fast conducting A fibers begin to produce SP. Thus, showing that nodose ganglia neurons contribute to the tachykinergic innervation of inflamed airways is consistent with the recent findings of Fischer and coworkers who demonstrated that within 1 d after allergen exposure preprotachykinin gene expression and SP production are elevated in nodose ganglia neurons (9).

Exposing the trachea to NGF also caused a qualitative change in the nature of jugular neurons expressing SP. Consistent with our previous studies (2), the conduction velocity analysis of guinea pig airway sensory nerves arising from the jugular ganglia was divided equally in number between C fibers and A fibers. In saline-treated animals the airway- labeled SP-positive neurons were uniformly found in small- diameter cells that were not NF-IR. This is consistent with the conclusion that these neurons belong to the capsaicin-sensitive C-fiber class (2). As with the nodose ganglia, after NGF treatment there was a phenotypic switch in the airway-specific tachykinergic population of neurons in the jugular ganglia such that approximately 30% of the large-diameter, NF-positive (presumably A-fiber) neurons projecting fibers to the trachea began to express SP-immunoreactivity.

There are substantive implications to a qualitative change in the tachykinergic innervation of the airways from jugular C fibers to nodose and jugular A fibers. Unlike the jugular C-fiber nerve endings, nodose nerve endings in the guinea pig airway are very sensitive to mechanical stimulation (2). The neurons induced by NGF to produce SP may therefore be stimulated by mechanical forces that are not nociceptive, but may occur even during normal breathing. There is indirect evidence that the nodose fibers innervate the region of the airway beneath the basement membrane, whereas jugular nerve fibers extend into the epithelium (3). If the SP were to be released from the induced nodose fiber population, it would likely affect different cell populations than those released from the constitutive jugular C-fiber population. Perhaps more importantly, nodose fibers may project to different populations of secondary neurons in the central nervous system. As much as neuropeptides are transported to both the central and peripheral ends of the primary afferent fiber (10), the induction of neuropeptides in these mechanically sensitive neurons, may lead to exaggerated reflex response to innocuous stimuli. Finally, the data support the speculation that because nodose A fibers are unresponsive to capsaicin and bradykinin (2, 17), these conventional tools used to study the physiology of tachykinergic innervation may not be useful to study a significant component of that tachykinergic innervation induced by NGF.