www.elsevier.com/locate/brainres

Brain Research 1031

Research report

Innervation of the rat trachea by bilateral cholinergic projections from the

nucleus ambiguus and direct motor fibers from the cervical spinal cord:

a retrograde and anterograde tracer study

Yasuro Atojia,*, Dwi Liliek Kusindartaa,b, Nao Hamazakia, Akihisa Kanekoa

aLaboratory of Veterinary Anatomy, Faculty of Applied Biological Sciences, Gifu University, Yanagido 1-1, Gifu 501-1193, JapanbDepartment of Anatomy, Faculty of Veterinary Medicine, Gadjah Mada University, Yogyakarta 55281, Indonesia

Accepted 19 October 2004

Available online 2 December 2004

Abstract

A tract–tracer method was employed to examine the innervation of the rat trachea. Cholera toxin h subunit (CTB) was injected into the

following locations in separate groups of rats: (1) ventral trachea, (2) lateral trachea, (3) ventral trachea after the excision of the nodose

ganglion, and (4) ventral trachea after the transection of C1–C2 spinal nerves. CTB injection in the ventral trachea showed bilateral

labeling of neurons in the nucleus ambiguus (NA), medial subnucleus of the nucleus of the solitary nucleus, dorsal motor nucleus of the

vagus (DMV), and lamina IX of C1–C6. CTB injection in the lateral trachea showed significant ipsilateral predominance of neuronal

labeling in the NA and lamina IX of C1–C2 segments. CTB injection in rats after the excision of the nodose ganglion revealed no

labeling in the ipsilateral DMV and NA and a significant reduction of neuronal labeling in C1. CTB injection in rats after the transection

of C1–C2 spinal nerves showed a significant decrease in the number of labeled neurons in ipsilateral NA, C1, and C2 and no labeling of

fibers in C1–C2. The combination of retrograde fluorogold labeling and choline acetyltransferase (ChAT) immunostaining revealed that all

fluorogold-labeled neurons in the NA and lamina IX of C1–C2 colocalized with ChAT. The injection of biotinylated dextran amine in NA

produced labeling in axonal terminals on postganglionic neurons, but not in other regions of the trachea. Our findings indicate that the rat

trachea is innervated bilaterally by cholinergic motor neurons in NA and C1–C2, while those traveling through the spinal nerves project

directly to the trachea.

D 2004 Elsevier B.V. All rights reserved.

Theme: Motor systems and sensorimotor integration

Topic: Spinal cord and brainstem

Keywords: Airway; Nerve transection; Sensory innervation; Cholera toxin; Dextran amine

1. Introduction

Autonomic innervation of the airway includes the para-

sympathetic and sympathetic nervous systems, which

function to control smooth muscle tension, secretion of

mucosal glands, vascular tone, and tracheobronchial

reflexes [3,7]. The preganglionic parasympathetic neurons

0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.brainres.2004.10.032

* Corresponding author. Tel.: +81 58 293 2936; fax: +81 58 293 2840.

E-mail address: atozi@cc.gifu-u.ac.jp (Y. Atoji).

innervating the trachea originate from the dorsal motor

nucleus of the vagus (DMV) and nucleus ambiguus (NA)

[5,8]. Postganglionic parasympathetic neurons are located in

the peritracheal ganglion or submucosal plexus [1,9,11,14].

Sensory fibers that are distributed in the tracheal epithelium

project to the nucleus of the solitary tract (NTS) through

neurons located in the nodose or jugular ganglion.

The innervation pattern of vagal motor neurons to the

lung shows species differences: NA in the cat innervates the

lung contralaterally [8], whereas in the rat, it innervates the

lung bilaterally [12]. The contralateral pathway in bilateral

(2005) 90–100

Y. Atoji et al. / Brain Research 1031 (2005) 90–100 91

innervation undergoes decussation inside the thorax in the

rat [12]. In the trachea, however, the issue of whether the

trachea is innervated bilaterally or contralaterally remains to

be resolved, although both DMV and NA participate in

tracheal innervation [5].

A tract-tracing study using cholera toxin h subunit

(CTB) showed that neurons in the dorsomedial part of the

ventral horn in C1 and C2 participate in the innervation of

the trachea in the dog, ferret, and rat, in addition to NA and

DMV [5]. Furthermore, the authors found CTB-labeled

fibers in the ventral funiculus of the upper cervical segments

in the dog and in the cervical dorsal horn in the ferret. The

study suggested that the upper cervical spinal cord

participates partially in the control of the trachea.

Sensory or parasympathetic nerve fibers are distributed

in the epithelium or lamina submucosa of the trachea of

the ferret and rat [4,9,13]. Perez-Fontan and Velloff [11]

reported that the injection of CTB in the tracheal lumen

resulted in labeling neurons in NA and DMV by trans-

epithelial transport to intra- or subepithelial space, where

labeled vagal motor neurons project their axons in the rat.

The authors hypothesized that vagal motor neurons in

DMV and NA project their axons directly to the

epithelium or submucosa without interposition of intrinsic

neurons, which are located in the peritracheal para-

sympathetic ganglion. This hypothesis is contradictory to

the classic concept of the airway parasympathetic trans-

mission; preganglionic vagal neurons synapse on post-

ganglionic neurons in the local ganglion, and, in turn, the

latter exert their effects on smooth muscles, glandular

cells, and epithelial cells [7,10]. However, their experiment

was carried out under the presence of the classic

parasympathetic rely [11]. Therefore, the hypothesis of a

direct vagal motor innervation to the trachea remains to be

confirmed.

The aims of the present study were to determine whether

NA and DMV innervate trachea bilaterally and whether the

cervical motor neurons directly innervate the trachea.

Experiments involved tract tracing using CTB and denerva-

tion of the vagus nerve and spinal nerves. In the latter

experiment, we chose the cervical spinal cord, where

autonomic innervation of both sympathetic and parasympa-

thetic nerves is absent [2]. CTB was injected in the tracheal

wall because previous studies reported that retrograde-

labeled neurons were more numerous in NA and DMV

than those obtained by injection into the tracheal lumen

[5,11].

2. Materials and methods

A total of 52 Wistar rats (weight 250–300 g, 10–16

weeks, of both sexes) was used in the present study. The

experimental protocol was approved by the Ethics Review

Committee for Animal Experimentation of Faculty of

Agriculture, Gifu University.

2.1. Injection of retrograde tracers in the trachea

Experiments consisted of five groups to clarify bilateral

innervation of the medulla to the trachea and innervation of

motor neurons in the cervical spinal cord to the trachea. Rats

were anesthetized by intraperitoneal injection of sodium

pentobarbital (25 mg/kg body weight). Under sterile

conditions, the cervical trachea was exposed. In rats of

group 1 (n=5), 0.5% CTB (low salt, List Biological

Laboratories, USA) in phosphate-buffered saline (PBS)

was injected in the ventral midline of the trachea in the

5th to 14th intercartilagenous spaces, using a glass micro-

pipette (a total of 10 injections of 400 nl each, total

volume=4 Al). In rats of group 2 (n=5), the injection

protocol was similar to that used for group 1, except that the

injections were made in the left or right side of the trachea.

In rats of group 3 (n=5), the left or right nodose ganglion

was excised, followed 7–10 days later by tracheal injections

under anesthesia of the same toxin solution used for group

1, at the same dose and in the same anatomical site. In rats

of group 4 (n=5), 2 mm of the right C1 and C2 spinal

nerves, just lateral to the intervertebrate foramen, was cut,

followed 7–10 days later by tracheal injections under

anesthesia of the same toxin solution used for group 1, at

the same dose and in the same anatomical site. In rats of

group 5, 0.5% fluorogold (Biotum, USA) in distilled water

was injected in the ventral midline of the trachea in the 5th

to 14th intercartilagenous spaces using a glass micropipette

(a total of 10 injections of 400 nl each, total volume=4 Al) todetect colocalization with choline acetyltransferase (ChAT),

similar to group 1.

Three controls of injections were done to examine

whether tracers injected in the trachea were leaked into

neighboring tissues, i.e., the m. sternohyoideus, which

covers the ventral surface of the trachea. In rats of control

1 (n=3), after the m. sternohyoideus of both sides were

removed between near origin and insertion, CTB was

injected in the ventral midline of the trachea, in the 5th to

14th intercartilagenous spaces (a total of 10 injections of

400 nl each, total volume=4 Al). In rats of control 2 (n=3),

CTB was injected in the m. sternohyoideus of both sides

(each side received a total of 10 injections of 400 nl each,

total volume=4 Al). In rats of control 3 (n=3), fluorogold

was injected in the ventral midline of the trachea in the 5th

to 14th intercartilagenous spaces (a total of 10 injections

of 400 nl each, total volume=4 Al), while CTB was

injected in the m. sternohyoideus of both sides (each side

received a total of 10 injections of 400 nl each, total

volume=4 Al).On the 4th day after the injection of CTB or fluorogold,

all rats were anesthetized with sodium pentobarbital (50 mg/

kg, intraperitoneal injection), perfused with Ringer’s sol-

ution, followed by 4% formaldehyde in 0.1 M phosphate

buffer at pH 7.4. The brainstem and cervical spinal cord

were dissected out and postfixed in the same fixative for 2–3

days. Specimens were transferred to 30% sucrose in PBS at

Fig. 1. Diffusion of CTB (A) and fluorogold (B) in the trachea after injections in the ventral midline. Arrows indicate the diffusion of tracers. Scale bars=1 mm.

Y. Atoji et al. / Brain Research 1031 (2005) 90–10092

4 8C. Brainstems and spinal cords were cut at 50-Amthickness on a cryostat.

2.2. Injection of anterograde tracer in the nucleus ambiguus

Rats were anesthetized with sodium pentobarbital (50

mg/kg, intraperitoneal injection) and placed in a stereotaxic

Fig. 2. Retrograde CTB labeling in the medulla. (A) Schematic illustration of half

Enlargement of the box in Panel (A). CTB-labeled neurons are seen in the dorsal

Am. (C) Schematic illustration of half medulla showing three labeled nuclei at +600

the nucleus of the solitary tract. Scale bar=100 Am. (E) CTB-labeled neurons in th

neurons are multipolar or bipolar in shape and are closely packed in NAd. Scale

instrument (Narishige, Tokyo, Japan). Each animal received

an injection of 10% biotinylated dextran anime (BDA,

10,000 molecular weight; Molecular Probes, Eugene, OR)

on bilateral NA. BDA in PBS was iontophoretically injected

through glass micropipettes (outer diameter 10–15 Am; 5 AAcurrent; 5 s on, 5 s off for 15 min). Following survival for 1–

3 weeks, the animals were anesthetized with sodium

medulla showing the nucleus ambiguus at +800 Am rostral to the obex. (B)

(NAd) and ventral divisions (NAv) of the nucleus ambiguus. Scale bar=200

Am. (D) Abundant CTB-labeled axon terminals in the medial subnucleus of

e dorsal motor nucleus of vagus nerve. Scale bar=50 Am. (F) Labeled large

bar=100 Am.

Y. Atoji et al. / Brain Research 1031 (2005) 90–100 93

pentobarbital (50 mg/kg) and perfused with Ringer’s

solution, followed by 4% paraformaldehyde in 0.1 M

phosphate buffer at pH 7.4. Brains and tracheas were

removed, postfixed in the same fixative for 2–3 days, and

transferred to 30% sucrose in PBS at 4 8C. The medulla was

transversely sectioned at 50 Am on a cryostat, whereas the

trachea was sectioned at 50 Am, and the sections were

mounted on glass slides.

2.3. Immunohistochemical staining

Sections of rats of groups 1–4 were stained with an

antibody against CTB. Sections of the medulla and cervical

spinal cord were rinsed in PBS and were treated with 50%

methanol containing 0.3% H2O2 for 30 min. After rinsing in

PBS, they were incubated with 1% nonimmune rabbit serum

for 1 h. The specimens were then incubated with a solution

containing goat anti-CTB antibody (1:60,000; List Bio-

logical Laboratories), 0.3% Triton X-100, and 1% rabbit

serum for 2 days at 4 8C, followed by incubation with

biotinylated rabbit antigoat IgG (1:500; Sigma, St. Louis,

Fig. 3. Histograms showing the distribution of labeled neurons in the nucleus ambi

solitary tract (NTS), and dorsal motor nucleus of the vagus nerve (DMV) after

continuity of two regions of NA (+1800 to �1800 Am) and cervical spinal cord (C

and ventral divisions. (B) Histogram shows continuity of two nuclei of NTS (+1800

of labeled neurons in four consecutive sections (n=5). The numbers of labeled ne

labeled neurons was enumerated in all sections (40–50 sections) of each segmen

(n=5). Data are meanFS.D. SC: spinal cord.

MO) for 1 h at room temperature, and finally incubated in

avidin–biotin–peroxidase complex (ABC, Elite ABC kit;

Vector Laboratories, Burlingame, CA) for 1 h at room

temperature. ABC was visualized by incubation in 0.05 M

Tris–HCl buffer (pH 7.4) containing 3, 3V-diaminobenzidine

tetrahydrochloride (DAB) and H2O2. Sections were

mounted on glass slides precoated with gelatin and chrome

alum, dehydrated in a series of ethanol solutions, cleared in

xylene, and coverslipped.

Sections of the brainstem and spinal cord of group 5 were

stained with goat anti-ChAT. Sections were first treated with

50% methanol containing 0.3% H2O2 for 30 min. Then,

they were incubated with 1% nonimmune donkey serum for

30 min. After washing in PBS, sections were incubated with

goat anti-ChAT antibody (1:400; Chemicon International,

Temecula, CA) overnight at 4 8C. The sections were

incubated with antigoat IgG conjugated with tetramethylr-

hodamine isothiocyanate (TRITC, 1:100; Jackson Immu-

noresearch Laboratories, West Grove, PA) for 2 h at room

temperature in the dark. After washing in PBS, the sections

were mounted on glass slides, coverslipped with carbonate-

guus (NA), cervical spinal cord, the medial subnucleus of the nucleus of the

CTB injection in the ventral midline of the trachea. (A) Histogram shows

1 to C6). The number of labeled neurons in NA indicates the sum of dorsal

to +1000 Am) and DMV (+800 to �2000 Am). Each bar expresses the total

urons in C1–C6 bars were calculated as follows: First, the total number of

t, divided as equivalent to four sections, and mean values were calculated

Y. Atoji et al. / Brain Research 1031 (2005) 90–10094

buffered glycerol, and examined with a fluorescence

microscope.

Sections of the medulla and trachea in the BDA-injected

rats were treated with 50% methanol containing 0.3% H2O2

for 30 min. After washing in PBS, they were incubated with

ABC (Vector Laboratories) in PBS containing 0.3% Triton

X-100 for 2 h at room temperature, followed by incubation

in 0.05 M Tris–HCl buffer (pH 7.4) containing DAB and

H2O2 in 0.1 M Tris–HCl buffer at pH 7.4. Tracheal sections

mounted on glass slides and floating sections were stained

as described above.

2.4. Data analysis

The numbers of labeled neurons in the medulla and

cervical spinal cord of groups 1–4 (n=5) were counted.

Differences between groups were examined for statistical

significance using Student’s t-test. Statistical significance

was established as Pb0.05 or 0.01, and all data were

expressed as meanFS.D.

3. Results

3.1. Spread of tracers in the trachea

In cases of CTB injections, the tracer was not found in

the trachea 4 days after the injections. Therefore, the

diffusion of the tracer was examined in 30 min and 2, 24,

and 48 h after the injections in the ventral midline of the

trachea. Thirty minutes after injection, CTB was seen in the

Fig. 4. Retrograde CTB labeling in the cervical spinal cord. (A) Schematic ill

Enlargement of the box in (A). Abundant labeled axons or terminals in the central c

nucleus. Scale bar=50 Am. (D) Many labeled neurons in the medial part of lamin

ventral part of the trachea (Fig. 1A). Two and 24 h after

injections, CTB extended into the ventral half of the trachea,

but 48 h after the injections, it was not yet detected in the

trachea. In cases of fluorogold injections, the tracer could be

still seen in the trachea 4 days after injections. In the midline

injection, fluorogold was found mainly in the ventral part

(Fig. 1B), while in the lateral injection, the tracer was

restricted to the ipsilateral side.

3.2. Injection of CTB and fluorogold

The injection of CTB in the ventral midline of the

tracheal wall (group 1) showed labeling neurons in the

dorsal (NAd) and ventral (NAv) divisions of the nucleus

ambiguus, DMV, and rostral part of the medial subnucleus

of NTS in the medulla oblongata (Fig. 2A–C, E, and F). The

number of labeled neurons was largest in NA among the

three nuclei (Fig. 3A and B). Labeled neurons in NAd were

largely multipolar and extended rostrocaudally from +1.8

(the obex is 0 mm) to �0.4 mm, whereas multipolar labeled

neurons in NAv were found at the levels of +1.4 to +0.8

mm. The densest of labeled neurons in the whole NA was

found in +1.20 mm rostral to the obex (Fig. 3A). Labeled

neurons were found in the rostral part of the medial

subnucleus of NTS, at +1.8 to +1.0 mm. Round or bipolar

small labeled neurons were found in the DMV, which was

localized in +0.8 to �2.0 mm. Labeled fibers in the NTS

were seen in the medial, ventrolateral, and commissural

subnuclei at +1.6 to �2.2 mm (Fig. 2D). The densest

labeled fibers in the NTS were localized in the medial

subnucleus at levels of +0.4 to �0.4 mm. In the cervical

ustration of half spinal cord showing the orientation of B–D at C1. (B)

ervical nucleus. Scale bar=100 Am. (C) A labeled neuron in the intermediate

a IX. Scale bar=100 Am. CC: central canal.

Fig. 5. Histograms showing the numbers of labeled neurons. (A) Number of labeled neurons when CTB was injected in the left lateral side of the trachea. (B)

Number of labeled neurons when CTB was injected after the excision of the nodose ganglion. (C) Number of labeled neurons when CTB was injected after

transection of C1–C2 spinal nerves. Each bar in the dorsal motor nucleus of the vagus nerve (DMV), nucleus ambiguus (NA), and C1–C6 expresses the total

number of labeled neurons in the nuclei and segments (n=5). Statistical analysis was not performed from C3 to C6 because of the low number of labeled

neurons. Data are meanFS.D. **Pb0.01, *Pb0.05.

Y. Atoji et al. / Brain Research 1031 (2005) 90–100 95

Fig. 6. CTB labeling after the excision of the right nodose ganglion (B–E) and transaction of the right C1–C2 spinal nerves (G–J). (A) Schematic illustration of

the medulla showing the orientation of B–E at +800 Am. (B and C) Enlargement of the respectively labeled boxes in Panel (A). Many CTB-labeled neurons are

seen in the left nucleus ambiguus (NA; B), but not in the right NA (Panel C). (D and E) CTB-labeled fibers are numerous in the left nucleus of the solitary tract

(NTS; D), but not in the right side of NTS (E). (F) Schematic illustration showing the orientation of G–J in C1. (G and H) Enlargement of the respectively

labeled boxes in Panel (F). CTB-labeled neurons are seen in the left lamina IX (G), but not in the right lamina IX (H). CTB-labeled fibers are found in left

central cervical nucleus (CeC; I), but not in the right CeC (J). Scale bars=100 Am in Panels (B) and (C); 50 Am in Panels (D), (E), (I), and (J); 25 Am in Panels

(G) and (H).

Y. Atoji et al. / Brain Research 1031 (2005) 90–10096

spinal cord, CTB-labeled neurons were detected mainly in

the medial part of lamina IX, from C1 to C6, and a few in

the lateral part of lamina IX (Fig. 4A and D). In some cases,

one or two labeled neurons were seen in C7. Labeled

neurons were also seen in the intermediate nucleus (Fig. 4C)

Fig. 7. Colocalization of retrogradely labeled fluorogold (FG) and immunohisto

fluorogold-labeled neurons in the nucleus ambiguus are ChAT immunoreactive (arr

of the vagus nerve are ChAT immunoreactive (arrows), and the others are negativ

medial part of lamina IX are ChAT immunoreactive (arrows). C1. Scale bar=50 A

and medial edge of lamina VII in C1. The number of labeled

neurons was densest in C1 and decreased as segments ran

caudally. Labeled varicose fibers were seen in an area

covering the central cervical nucleus and medial edge of

lamina IV in C1 to C3 (Fig. 4B).

chemical staining of choline acetyltransferase (ChAT). (A and B) Three

ows). (C and D) Two fluorogold-labeled neurons in the dorsal motor neuron

e for ChAT (arrowheads). (E and F) Two fluorogold-labeled neurons in the

m.

Y. Atoji et al. / Brain Research 1031 (2005) 90–100 97

When CBT was injected in the ventral midline of the

tracheal wall (group 1), the distribution pattern and

number of labeled neurons in NA, DMV, and C1–C6

Fig. 8. Schematic illustration of the rostrocaudal extent (C1 to C6) of labeling

column, A–F), in the ventral midline of the trachea after removing the m. sternohyoi

R), respectively. Dots indicate the relative differences in distribution but do not rep

labeled axons. Scale bar=1 mm.

were almost identical on both sides (Fig. 3A and B).

However, when CTB was injected in the left side of the

tracheal wall (group 2), the number of labeled neurons in

following injections of CTB in the ventral midline of the trachea (left

deus (middle column, G–L), and in the m. sternohyoidues (right column, M–

resent real numbers. Hatched area indicates the distribution of anterogradely

Fig. 9. Labeled neurons in the medial part of lamina IX of C1 after fluorogold injection in the trachea and CTB injection in the m. sternohyoideus. Arrows

indicate colocalization of fluorogold and CTB. Arrowheads show a fluorogold-positive, CTB-negative neuron. Scale bar=25 Am.

Y. Atoji et al. / Brain Research 1031 (2005) 90–10098

the left side of the NA (249.3F67.4) was significantly

higher than in the right side (141.3F63.9, n=5, Pb0.01),

in the left side of C1 (170.2F57.7) than in the right

(87.5F33.4, n=5, Pb0.01), and in the left side of C2

(39.8F27.9) than in the right (15.7F14.6, n=5, Pb0.05),

but not in DMV (Fig. 5A). Similar differences were noted

when CTB was injected in the right side of the trachea

(data not shown). When CTB was injected in the ventral

midline of the trachea wall in rats, after the removal of the

right nodose ganglion (group 3), no labeled neurons were

found in the NA and DMV of the ipsilateral side (Figs. 5B

and 6A–C), and the number of neurons in C1 decreased

significantly in the transected side (left: 117.8F57.5; right:

57.0F25.7; n=5, Pb0.05). There were also no labeled

terminal fibers in the right NTS (Fig. 6D and E).

Differences were also detected in rats with excised left

nodose ganglion (data not shown). When CTB was

injected in the ventral midline of the tracheal wall, after

the transection of the right C1–C2 spinal nerves (group 4),

labeled neurons were significantly fewer in the ipsilateral

NA (left: 135.6F26.9; right: 81.2F28.7; n=5, Pb0.01), C1

(left: 117.2F42.5; right: 29.6F24.0; n=5, Pb0.01), and C2

(left: 20.8F8.3; right: 2.0F1.2; n=5, Pb0.01) than in the

contralateral sides (Figs. 5C and 6F–H). Labeled fibers in

the right central cervical nucleus disappeared in C1–C2

(Fig. 6I and J).

In group 5, fluorogold-labeled neurons were only found

in the NA, DMV, and lamina IX of the cervical spinal

cord in the central nervous system (data not shown).

Fig. 10. BDA labeling in the nucleus ambiguus (NA) and trachea. (A) Injection si

bar=1 mm. (B) BDA-labeled axon terminals are seen on postganglionic neurons

ChAT was colocalized in all fluorogold-labeled neurons in

the NA and lamina IX of C1–C2 and some neurons in

DMV (Fig. 7A–F).

3.3. Control injections

In control 1, removing the m. sternohyoideus, the

distribution of labeled neurons and varicose fibers (Fig.

8A–F) did not show a difference compared with those of

group 1 (Fig. 8G–L). In control 2, injecting in the m.

sternohyoideus, no labeling was seen in the medulla

oblongata, including NA, DMV, or NTS. In the cervical

spinal cord, labeled neurons were found in most lamina IX

from C1 to C6 (Fig. 8M–R), but labeled varicose fibers were

not seen in all segments. In control 3, the percentage of

colocalization with CTB and fluorogold was 37.2% of

labeled neurons in the medial part of lamina IX of C1 (a

total 239 labeled neurons in three rats; Fig. 9A and B), and

the single labeling of fluorogold and CTB showed 41.9%

and 20.9%, respectively.

3.4. Injection of BDA in NA

When BDA was injected in the rostral part of NA,

labeled axon terminals were frequently found on cell bodies

in the peritracheal ganglion adjoining the smooth muscle

layer (Fig. 10A and B). However, no labeled axon terminals

or axons were detected in the lateral or ventral wall of the

trachea.

te of BDA in the right NA is at +1.2 mm rostral to the obex (arrow). Scale

in a peritracheal ganglion. SM: smooth muscle layer. Scale bar=25 Am.

Fig. 11. Schematic representation of the neural pathways involved in the

control of the trachea. The classic routes of parasympathetic and sensory

innervation are drawn by thin lines. In the current study, two distinct

pathways (thick lines) of motor neurons innervating the trachea originate in

NA and C1–C2. Nerve terminals of sensory neurons in nodose and C1–C2

dorsal root ganglia (DRG) project to C1 and C2 segments (thick lines).

DMV: dorsal motor nucleus of the vagus nerve; NA: nucleus ambiguus;

NTS: nucleus of the solitary tract; C1–C2: first and second segments of the

cervical spinal cord.

Y. Atoji et al. / Brain Research 1031 (2005) 90–100 99

4. Discussion

The present study demonstrated the bilateral innervation

of NA and C1–C2 to the trachea and the direct innervation

of motor neurons of the cervical spinal cord to the trachea in

the rat. The retrograde labeling of motor neurons in DMV

and NA and the axonal labeling in NTS in the medulla is in

good agreement with those of CTB tract-tracing studies

[5,6]. The combination of fluorogold labeling and ChAT

immunostaining showed that ChAT coexists on all fluo-

rogold-labeled neurons in the NA and lamina IX of the

spinal cord and some neurons in DMV. At least all motor

neurons innervating the trachea in NA and cervical spinal

cord are cholinergic.

4.1. Bilateral innervation of the trachea

Using cervical vagotomy and CTB or fluorogold injec-

tion, Perez-Fontan et al. [12] found that each NA innervates

the lung bilaterally in the rat, and the fiber pathway from the

NA to a contralateral lung crosses midline inside the thorax.

In the present study, CTB injection in the ventral midline of

the tracheal wall showed no significant difference in the

distribution of labeled neurons in the NA, DMV, and cervical

spinal cord, but CTB injection in the lateral wall showed not

only bilateral labeling but also significant predominance of

neuronal labeling in the injected side of NA, C1, and C2. The

evidence indicates that a single side of NA, C1, and C2

innervates both sides of the tracheal wall, as well as the lung.

CTB injection in rats after the removal of the nodose ganglion

(group 3) revealed no labeling of neurons in the vagotomized

side of NA and DMV. This finding indicates that CTB-

labeled neurons in the NA and DMV do not decussate in the

central region proximal to the nodose ganglion, but do in the

peripheral region of the vagus nerve. Therefore, it is likely

that the recurrent nerve containing neurons in the NA and

DMV, which innervate contralateral region of the trachea,

decussate inside the thorax as those in NA reach the lung

[12]. On the other hand, the number of labeled neurons in

vagotomized C1 was reduced to approximately half of that on

the intact side. This indicates that C1 neurons innervating the

trachea through the vagus nerve bilaterally decussate in the

central nervous region, i.e., in C1.

4.2. Direct, central innervation of the trachea

In the present study, the injection of CTB in the ventral

wall of the trachea resulted in the labeling of neurons of

lamina IX of C1–C6 and nerve fibers in the central cervical

nucleus of C1–C3. Labeled neurons in the cervical spinal

cord exhibited a wider distribution in the present study than

did that of Haxhiu and Loewy [5], who found neuronal

labeling in C1 and C2 of the dog, ferret, and rat. In control

3, the medial part of lamina IX of C1 contained three types

of labeled neurons: fluorogold-labeled neurons to the

trachea (41.9%), CTB-labeled neurons to the m. sterno-

hyoideus (20.9%), and double-labeled neurons to both the

trachea and m. sternohyoideus (37.2%). The percentage of

neurons innervating the trachea was 79.1% of all labeled

neurons. This shows the presence of a direct projection from

the medial part of lamina IX of C1 to the trachea, in addition

to the motor neurons to the m. sternohyoideus. Labeled

fibers were seen in the ventral funiculus of the upper

cervical segments in the dog and in the cervical dorsal horn

of the ferret, but no labeling was observed in the cervical

dorsal horn in the rat [5]. However, these fiber labeling

patterns are in sharp contrast to those in the present study, in

which CTB-labeled fibers were located mainly in the central

cervical nucleus but not in the dorsal horn or ventral

funiculus. The evidence that CTB-labeled fibers in the

central cervical nucleus of C1–C2 disappeared ipsilaterally

following the severance of C1–C2 spinal nerves suggests

that CTB-labeled fibers in the central cervical nucleus are

centrifugal axons of sensory neurons in C1–C2 dorsal root

ganglia, with projections peripherally to the trachea (Fig. 8).

The injection of CTB in the intercartilagenous space

labeled the neurons in the NA and DMV of the dog, ferret,

and rat [5]. The intraluminal injection of CTB into the

trachea also produced the labeling of neurons in the NA and

DMV in the rat [11]. The authors in the latter study suggested

the presence of direct innervation of the trachea by vagal

motor neurons. However, in these studies, possible labeling

by transganglionic transport via the classic parasympathetic

pathway could not be excluded because the classic para-

sympathetic pathway was still intact anatomically and

functionally when CTB was injected in the tracheal wall or

tracheal lumen. In the present study, the injection of CTB in

Y. Atoji et al. / Brain Research 1031 (2005) 90–100100

the ventral wall of the trachea after cutting the right spinal

nerves of C1 and C2 resulted in a significant reduction of the

numbers of labeled neurons in the right lamina IX. The

reduction of the number of labeled neurons in the cut side

indicates the participation of cervical motor neurons in direct

innervation of the trachea (Fig. 8) because no preganglionic

neurons of either parasympathetic or sympathetic nerve are

present in the cervical spinal cord [2]. The transection of the

spinal nerves in the present study also revealed significant

reduction of the number of CTB-labeled neurons in ipsi-

lateral NA. The evidence indicates the presence of descend-

ing neurons that run caudally in the lower medulla, enter the

upper cervical spinal cord, and go through C1–C2 spinal

nerves. In contrast, a significant reduction of the number of

CTB-labeled neurons in C1 was detected following the

excision of the nodose ganglion. The two experiments of

nerve transection indicate that a small population of motor

neurons in the NA reaches the trachea via the C1 and C2

spinal nerves, and inversely, those in C1 travel to the trachea

via the vagus nerve. That is, motor neurons located in the

transitional region between the medulla and upper cervical

spinal cord have two routes to reach the trachea. Further-

more, neurons in the NA, C1, and C2 that innervate the

trachea bilaterally via the spinal nerves decussate in a central

region, but not in a peripheral one (Fig. 11).

In the present study, BDA injection in the rostral NA,

where many neurons projected to the trachea (group 1),

revealed the labeling of axon terminals on postganglionic

neurons in the peritracheal ganglion, but no labeling in the

nerve fibers or axon terminals of other regions of the

tracheal wall. These results indicate that motor neurons in

the NA terminate on ganglionic neurons and suggest the

absence of direct innervation of the trachea without the

interposition of parasympathetic postganglionic neurons.

Therefore, it is likely that retrograde neuronal labeling in the

NA, DMV, and NTS after CTB injection in the trachea is

due to transganglionic transport. This is probably supported

by the finding that intrinsic tracheal ganglionic neurons

were labeled in addition to the NA, DMV, and cervical

spinal cord when fluorogold was injected in the ventral wall

in the present study (data not shown).

References

[1] D.G. Baker, D.M. McDonald, C.B. Basbaum, R.A. Mitchell, The

architecture of nerves and ganglia of the ferret trachea as revealed by

acetylcholinesterase histochemistry, J. Comp. Neurol. 246 (1986)

513–526.

[2] P. Brodal, Peripheral autonomic nervous system, The Central

Nervous System, 2nd edition, Oxford University Press, New York,

1998, pp. 493–528.

[3] B.J. Canning, A. Fischer, Neural regulation of airway smooth muscle

tone, Respir. Physiol. 125 (2001) 113–127.

[4] R.D. Dey, B. Satterfield, J.B. Altemus, Innervation of tracheal

epithelium and smooth muscle by neurons in airway ganglia, Anat.

Rec. 254 (1999) 166–172.

[5] M.A. Haxhiu, A.D. Loewy, Central connections of the motor and

sensory vagal systems innervating the trachea, J. Auton. Nerv. Syst.

57 (1996) 49–56.

[6] M.A. Haxhiu, A.S.P. Jansen, N.S. Cherniack, A.D. Loewy, CNS

innervation of airway-related parasympathetic preganglionic neurons:

a transneuronal labeling study using pseudorabies virus, Brain Res.

618 (1993) 115–134.

[7] D. Jordan, Central nervous pathways and control of the airways,

Respir. Physiol. 125 (2001) 67–81.

[8] M. Kalia, M.M. Mesulam, Brain stem projections of sensory and

motor components of the vagus complex in the cat: II. Laryngeal,

tracheobronchial, pulmonary, cardiac, and gastrointestinal branches, J.

Comp. Neurol. 193 (1980) 467–508.

[9] D.W. Kusindarta, Y. Atoji, Y. Yamamoto, Nerve plexuses in the

trachea and extrapulmonary bronchi of the rat, Arch. Histol. Cytol. 67

(2004) 41–55.

[10] A.C. Myers, Transmission in autonomic ganglia, Respir. Physiol. 125

(2001) 99–111.

[11] J.J. Perez-Fontan, C.R. Velloff, Labeling of vagal motoneurons and

central afferents after injection of cholera toxin B into the airway

lumen, Am. J. Physiol., Lung Cell. Mol. Physiol. 280 (2001)

L152–L164.

[12] J.J. Perez-Fontan, C.T. Diec, C.R. Velloff, Bilateral distribution of

vagal motor and sensory nerve fibers in the rat’s lungs and

airways, Am. J. Physiol., Regul. Integr. Comp. Physiol. 279 (2000)

R713–R728.

[13] M. Terada, T. Iwanaga, H. Takahashi-Iwanaga, I. Adachi, M.

Arakawa, T. Fujita, Calcitonin gene-related peptide (CGRP)-immu-

noreactive nerves in the tracheal epithelium of rats: an immunohis-

tochemical study by means of whole mount preparations, Arch.

Histol. Cytol. 55 (1992) 219–233.

[14] Y. Yamamoto, T. Ootsuka, Y. Atoji, Y. Suzuki, Morphological and

quantitative study of the intrinsic nerve plexuses of the canine trachea

as revealed by immunohistochemical staining of protein gene product

9.5, Anat. Rec. 250 (1998) 438–447.