The origin of catecholaminergic nerve fibers in the subdiaphragmatic vagus nerve of rat

Journal of the Autonomic Nervous System 76 Ž1999. 108–117

The origin of catecholaminergic nerve fibers in the subdiaphragmatic vagus nerve of rat Ming Yang, Xishun Zhao, Richard R. Miselis

)

Department of Animal Biology and the Institute of Neurological Sciences, School of Veterinary Medicine, UniÕersity of PennsylÕania, 3800 Spruce Street, Philadelphia, PA 19104-6045, USA Received 1 August 1997; received in revised form 1 March 1999; accepted 1 March 1999

Abstract It is known that the vagus nerve contains catecholaminergic fibers. However, the origin of these fibers has not been systematically examined. In this study, we addressed this issue using retrograde tracing from the subdiaphragmatic vagus nerve combined with immunocytochemistry. The cervical and thoracic sympathetic trunk ganglia, the nodose ganglia and the dorsal motor nucleus of the vagus nerve were examined following injection of Fluoro-Gold or cholera toxin horseradish peroxidase conjugate into the trunks of the subdiaphragmatic vagus nerve of rats. Numerous retrogradely labeled neurons were seen in the nodose ganglion and the dorsal motor nucleus of the vagus nerve. Very few labeled neurons were found in the sympathetic ganglia Žless than 0.06% of the neurons in either superior cervical ganglion or cervicothoracic ganglion were retrogradely labeled.. Double labeling with immunofluoresence for catecholamine synthesizing enzymes revealed that: Ž1. 92% of all Fluoro-Gold retrogradely labeled tyrosine hydroxylase immunoreactive neurons were found in parasympathetic sources Ž75% in the dorsal motor nucleus of the vagus nerve and 17% in the nodose ganglia., and only 8% in the cervicothoracic sympathetic ganglia; Ž2. 12% of the retrogradely labeled catecholaminergic neurons in the dorsal motor nucleus of the vagus nerve were also dopamine-b-hydroxylase immunopositive neurons; Ž3. 70% of the retrogradely labeled neurons in the sympathetic ganglia were tyrosine hydroxylase immunopositive and 54% of these catecholaminergic neurons contained dopamine-bhydroxylase, while 30% of the retrogradely labeled neurons were non-catecholaminergic neurons. These results indicate that catecholaminergic fibers in the abdominal vagus nerve are primarily dopaminergic and of parasympathetic origin, and that only an extremely small number of these fibers, mostly noradrenergic in nature, arise from postganglionic sympathetic neurons. q 1999 Published by Elsevier Science B.V. All rights reserved. Keywords: Superior cervical ganglion; Cervicothoracic ganglion; Nodose ganglion; Dorsal motor nucleus of the vagus nerve; Vagus nerve; Catecholamines

1. Introduction Early histofluorescent studies reported the presence of adrenaline nerve fibers in the vagus nerve trunks of the rat ¨ ŽOshumi et al., 1974., cat ŽMuryobayashi et al., 1968; Nielsen et al., 1969; Liedberg et al., 1973., dog ŽMuryobayashi et al., 1968; Ahlman et al., 1979., and human ŽLundberg et al., 1976.. A reduction of histofluorescent fibers in the vagus nerve was observed following ligation of the sympathetic trunk at either cervical or thoracic levels, suggesting a sympathetic origin to some of these fibers ŽMuryobayashi et al., 1968; Nielsen et al., ¨ 1969; Liedberg et al., 1973; Oshumi et al., 1974; Lundberg )

Corresponding author. Tel.: q1-215-898-6781; fax: q1-215-8989923.

et al., 1976; Ahlman et al., 1979.. Subsequent studies demonstrated that a subset of motor neurons in the dorsal motor nucleus of the vagus nerve ŽDMV. of the rat contain tyrosine hydroxylase ŽTH. ŽKalia et al., 1984; Sawchenko et al., 1987; Tayo and Williams, 1988; Ruggiero et al., 1993., dopamine-b-hydroxylase ŽDbH. ŽRitchie et al., 1982; Kalia et al., 1984; Gwyn et al., 1985; Ruggiero et al., 1993. or phenylethanolamine N-methyl transferase ŽPNMT. ŽKalia et al., 1984; Ruggiero et al., 1993.. In addition, some sensory neurons in the nodose ganglion ŽNG. of the rat contain TH ŽKatz et al., 1983, 1987; Helke and Niederer, 1990; Helke and Rabchevsky, 1991; Kummer et al., 1993.. These data suggest that catecholaminergic nerve fibers in the vagus nerve originate from parasympathetic neurons. However, the latter studies did not exclude the possibility of some catecholaminergic fibers of

0165-1838r99r$ - see front matter q 1999 Published by Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 1 8 3 8 Ž 9 9 . 0 0 0 1 4 - 4

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

sympathetic origin in the vagus nerve. To our knowledge, the subdiaphragmatic vagus nerve has not been injected directly with neuroanatomical tracers to unambiguously label putative sympathetic parent cell bodies. Another compelling reason to reinvestigate the origin of sympathetic fibers in the vagus nerve is that bilateral subdiaphragmatic vagotomy is a frequent manipulation in many behavioral, physiological and neuroanatomical studies where it is always presumed that vagotomy is a purely parasympathetic disruption. In the present study we, first, examined quantitatively the sympathetic origin of fibers in the subdiaphragmatic vagus nerve and, second, determined the proportion of vagal catecholaminergic fibers attributable to sympathetic and parasympathetic sources. 2. Methods Adult male Sprague–Dawley rats Ž250–350 g. were anesthetized with a mixture of Ketamine and Xylazine Ž10:7.5, 1.5 mlrkg, i.p... A midline ventral abdominal incision was made to access abdominal viscera. The ventral andror dorsal trunks of the subdiaphragmatic vagus nerve were carefully dissected free from the esophagus, and a small segment of the vagus trunk proximal to the hepatic branch was crushed for 30 s with a flat forceps. A glass micropipette with an outer tip diameter of 150 mm was inserted into each vagus nerve trunk 5–7 mm distal to the crushed site through a puncture in the nerve trunk previously created with a needle tip, and threaded up to the crush site where the tracer was slowly injected with a 5 ml Hamilton syringe. The tracers used were: 2 ml, 2–5% Fluoro-Gold ŽF-G, Fluorochrome Englewood, CO. in sterile saline Ž n s 20., 3.5 ml, 0.25% cholera toxin horseradish peroxidase conjugate ŽCT-HRP, n s 11. in 0.01 M phosphate buffered saline ŽPBS.. The following precautions were taken during the injection to prevent contamination of other nearby viscera: Ž1. placement of parafilm between the vagus nerve and the adjacent esophagus and covering nearby viscera with saline soaked gauze during injection; Ž2. ligation ŽNo. 4-0 suture. of the nerve trunk proximal to insertion site after injection to prevent leakage of tracer solutions from the puncture site; Ž3. at least three rinses Ž5 mlreach. of the abdomenal cavity with sterile saline to dilute any spillage. 2.1. Retrograde tracing 2.1.1. F-G injections Following 40–80 h of survival time, the rats were reanesthetized and perfused through the left cardiac ventricle with heparinized saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer ŽPB.. The superior cervical ganglion ŽSCG., cervicothoracic ganglion ŽCTG., a segment of the sympathetic trunk containing 8–10 thoracic sympathetic ganglia ŽTSG. caudal to the CTG, the

109

subdiaphragmatic vagus nerve, the NG and the caudal brainstem were dissected from the rat and postfixed overnight in the same fixative containing 20% sucrose. The ganglia were serially cut in 15 or 30 mm thick sections along their long axis in a cryostat at y188C, and mounted on glass slides. The caudal thoracic ganglia and trunk were coiled and mounted on a tissue stage to be cut simultaneously in horizontal sections. F-G retrogradely labeled neurons were counted with a Leitz Aristoplan fluorescence microscope using cube A with a band pass ŽBP. 360 nm excitation filter and a long pass ŽLP. 425 nm emission filter. 2.1.2. CT-HRP injections After surviving 40–72 h, the rats injected with CT-HRP were reanesthetized and perfused transcardially with 120 ml heparinized saline, 600 ml fixative Ž1% paraformaldehyde, 1.25% glutaraldehyde in 0.1 M PB., followed by 500 ml of 10% sucrose in 0.1 M PB. After perfusion, the brain, the NG, and the sympathetic ganglia of the cervicothoracic trunk, as described in the above section, were dissected and placed in 20% sucrose in 0.1 M PB overnight. Serial tissue sections Ž30 mm thick. of the various ganglia and brainstem were cut with a freezing microtome and processed with the tetramethylbenzidine ŽTMB. protocol ŽMesulam, 1982.. After the TMB reaction, the tissue sections were rinsed thoroughly in cold sodium acetate buffer ŽpH 3.3., mounted immediately on gelatin-coated slides, and stored at 48C overnight. Tissue sections were rapidly dehydrated in 100% ethanol for 20 s, cleared in xylene for 5 min, coverslipped with DPX mountant, and examined under light and dark-field microscopy. In the control experiments: Ž1. 2 ml of 2–5% F-G Ž n s 4. or 3.5 ml CT-HRP Ž n s 1. were slowly dripped on or underneath both vagal nerve trunks. After 5 min, the abdomen was flushed with saline as described above; Ž2. 12 or 15 ml of 2% F-G were injected directly into the abdominal cavity of rats Ž n s 2. without subsequent flushing. 2.2. Retrograde tracing and catecholamine immunocytochemistry Serial tissue sections were prepared for catecholamine labeling as follows: Ž1. three sets of serial Ževery 3rd. sections Ž15 mm thick. collected from sympathetic ganglia Ž n s 8., the NG Ž n s 10., and the subdiaphragmatic vagus nerve Ž n s 10. were mounted on gelatin coated slides and incubated in a drop Ž200 ml. of rabbit antibody ŽEugene Tech., NJ. to TH Ž1:200., DbH Ž1:150. or PNMT Ž1:200., respectively, at 258C for 20 h; Ž2. three sets of transverse sections Ž30 mm thick. of the lower medulla containing the DMV Ž n s 15. were floated in the same dilution of these antibodies Ž n s 5 for TH, DbH and PNMT, respectively., at 258C for 40 h. All sections were then incubated in donkey anti-rabbit serum ŽJackson Lab., NJ. tagged with FITC Ž1:40. at 258C for 3 h. After each step, all tissue

110

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

sections were rinsed with 0.01 M PBS Ž3 = 10 min.. F-G retrogradely labeled neurons were observed using cube A described above. Finally, catecholaminergic neurons labeled with FITC were observed using cube I3 with BP 470 nm excitation filter and LP 515 nm emission filter. In order to calculate the percentage of neurons in the SCG and the CTG that project through the subdiaphragmatic vagus nerve, we used stereological methods to obtain an unbiased estimate of the total number of neurons in these sympathetic ganglia. Five animals were anesthetized and perfused with heparinized saline followed by 4% paraformaldehyde. The sympathetic ganglia were dissected free and postfixed in the same fixative overnight before being transferred to 20% sucrose in 0.1 M PB ŽpH 7.4.. The ganglia were then serially sectioned at 15 mm in a cryostat, mounted on glass slides, and stained with thionin. For both the SCG and CTG, the best series of sections of either the left or right ganglia were used for stereological estimates of the total number of neurons. 2.2.1. Morphometrical analysis The two principal phases of applying stereology are the fractionator to obtain an estimate of the volume of the structure of interest and the optical disector to estimate neuronal density ŽCoggeshall and Lekan, 1996.. The total number of neurons in a ganglion Ž N . was determined by multiplying the volume of a ganglion Ž Vref . times the numerical density of the neurons Ž Nv , neurons per unit volume. in a ganglion, i.e. N s Vref Nv ŽCoggeshall, 1992.. Briefly, the volume of each ganglion Ž Vref . was estimated by multiplying the mean area of a set of sampling sections Ž asec . times the thickness of the sections Ž t; t s 15 mm. times the total number of sections through a ganglion Ž s ., i.e. Vref s asec ts ŽPakkenberg and Gundersen, 1988; Coggeshall, 1992.. Nv was determined by summing the neurons counted in each sampling volume Ž Q . and dividing by the sum of the sampling volumes Ž Vdis .. Vdis is the area of the sampling field Žalso called the disector. times the depth of the section through which nuclei were counted Žsee West et al., 1991.. Therefore, Nv s SQrSVdis . This was done using the set of sampling sections mentioned above to estimate Vref . In each sampling volume, the nucleus of the neuron was used as the counting object for each neuron. We used another stereological method, the physical disector ŽPakkenberg and Gundersen, 1988; Coggeshall and Lekan, 1996., to estimate the density of F-G labeled neurons for two reasons: Ž1. the small number of F-G labeled cells Žrange 0 to 73. observed in the SCG and CTG; and Ž2. the inability to discern organelles, such as nuclei, in the F-G labeled neurons. The number of F-G retrogradely labeled catecholaminergic neurons in the NG and DMV was counted in every other 30 mm thick section. The percentages of neurons singly labeled with F-G in the SCG or CTG were derived in this study. The mean number and the ratio of F-G retrogradely labeled catecholaminergic neurons in either sympathetic ganglia, SCG

and CTG, or in the parasympathetic structures ŽNG and DMV. were counted and calculated. 3. Results 3.1. Neurons single labeled with either F-G or CT-HRP in sympathetic ganglia 3.1.1. F-G labeled neurons Following the injection of F-G into both the ventral and dorsal trunks of the subdiaphragmatic vagus nerve, numerous strongly labeled cells were observed in the NG ŽFig. 1A., the DMV, and the compact nucleus ambiguus ŽNA.. Very few labeled neurons were found in the SCG ŽFig. 1B., the CTG ŽFig. 1C., and the TSG ŽFig. 1D.. F-G labeling in the sympathetic ganglia varied considerably. The mean number of labeled neurons was 6 " 1.88 S.E. Žrange, 0–22. in the SCG, 19 " 5.6 S.E. Žrange, 0–73. in the CTG, and 8 " 3.2 S.E. in each side for the series of TSG caudal to the CTG Žrange, 0–43.. Round and elongated labeled cells were located mainly in the ganglia ŽFig. 1D. with some fusiform shaped cells seen in the thin connecting segments of the sympathetic trunk. Retrogradely labeled neurons were round or elongated in shape, scattered, or distributed in small clusters at either one pole or near the margin of the middle portion of both SCG or the CTG ŽFig. 1B,C.. No labeled neurons could be found contralaterally in either NG or the various sympathetic ganglia of the animals that received only a single trunk tracer injection. In addition, labeled neurons did not occur in the spinal cord nor in any area of the brainstem other than the DMV and the NA in all vagus nerve injection cases. The average of our stereological estimates of the total number of neurons in the SCG from five rats was 36 706 " 16 749 S.E. The average in the CTG was 36 016 " 18 047 S.E. The range for the five different SCG was 30 159 to 47 250 and for the five CTG it was 33 708 to 40 500. The percentage of F-G retrogradely labeled neurons to the total number of neurons was 0.016% in the SCG and 0.053% in the CTG ŽTable 1.. The volume of the SCG and CTG varied from 0.52 to 1.1 mm3 and 0.57 to 0.9 mm3 , and the density of the neurons in these two ganglia was 35 – 63 neuronsrmm 3 and 39 – 57 neuronsrmm3 , respectively. 3.1.2. CT-HRP labeled neurons Following injection of CT-HRP into the trunks of the subdiaphragmatic vagus nerve, retrograde labeling in the DMV, NG ŽFig. 2A., and sympathetic ganglia was similar to that with the F-G injection. The distribution of labeled neurons, their size and shape were also similar to those seen with F-G labeling. Labeled neurons were seen in the SCG Žrange, 0–14, Fig. 2B., in the CTG Žrange, 0–53, Fig. 2C., and in the TSG ŽFig. 2D.. The mean number of labeled cells in the SCG was 2 " 1.05 S.E., and 13 " 4.27 S.E. in the CTG. The ratio of labeled neurons to the total

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

111

Fig. 1. The photomicrographs illustrate ganglionic labeling after F-G injections into both subdiaphragmatic vagal nerve trunks of the rat. A large number of F-G densely labeled neurons are seen in the nodose ganglion ŽA., while only a very few labeled cells Žarrows. can be seen in the superior cervical ganglion ŽB., the cervicothoracic ganglion ŽC. and the thoracic sympathetic ganglia ŽD.. SCG—superior cervical ganglion; CTG—cervicothoracic ganglion; NG—nodose ganglion; TSG—thoracic sympathetic ganglia. The calibration bars are equal to 200 mm.

number of neurons was 0.005% for the SCG and 0.038% for the CTG ŽTable 1.. 3.1.3. Control experiments Following the application of 2 ml of F-G directly on or beneath the subdiaphragmatic vagal nerve trunks, the num-

ber of labeled neurons in the NG was markedly reduced. In the sympathetic ganglia, fewer neurons Žrange, 0–2rper section. with noticeably fainter fluorescence ŽFig. 3C,D. were observed in comparison with that after direct vagal injection of F-G ŽFig. 3A,B.. In those animals subjected to F-G dropped into the abdomen, labeled neurons with fainter

Table 1 The mean number and percentage of labeled neurons to total neurons in the superior cervical ŽSCG. and the cervicothoracic ganglia ŽCTG. following F-G and CT-HRP injections Animal cases

20 11

Survival time Žh.

40–80 42–72

Tracers

F-G CT-HRP

Mean number of labeled neurons SCG

CTG

6 2

19 13

Total number of neurons SCG CTG

Percentage of labeling SCG Ž%. CTG Ž%.

36 706

0.016 0.005

36 016

0.053 0.038

112

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

Fig. 2. The photomicrographs illustrate the results from a CT-HRP injection into one subdiaphragmatic nerve trunk. Numerous labeled neurons are present in the NG ŽA., while very few labeled neurons occur in the SCG ŽB., the CTG ŽC. and the TSG ŽD.. In D, the big arrows point to CT-HRP labeled cells while small arrows point to artifact Žlabeled blood cells.. The calibration bars are equal to 200 mm.

fluorescence were observed bilaterally in the NG and the sympathetic ganglia of the cervicothoracic trunk, particularly in the SCG ŽFig. 3E,F.. No significant labeling, not even lighter labeling, was seen after dripping CT-HRP onto the vagal trunks. 3.2. Neurons double labeled for F-G and catecholamine enzymes in sympathetic and parasympathetic structures Neurons from serial sections of the SCG, CTG and TSG were immunostained for TH, DbH and PNMT. DbH immunoreactivity was more intense than TH immunoreactivity in these ganglia. There were no cells with PNMT immunoreactivity. Approximately 70% of the F-G retrogradely labeled neurons in the cervicothoracic divisions of the sympathetic ganglia had TH immunoreactivity. Seventy-seven percent of these retrogradely labeled TH posi-

tive neurons were also DbH positive, suggesting that most of these TH positive neurons are noradrenergic. Thirty percent of the retrogradely labeled neurons had neither TH nor DbH immunoreactivity. Neurons immunostaining for TH, DbH and PNMT were observed in the DMV. The mean number of TH immunopositive neurons was 380 Žrange, 371–385. and, of this number, 215 neurons were retrogradely labeled by F-G. A minority of these double labeled neurons was seen lying in the rostral extent of the DMV, while the majority were between the area postrema and spinomedullary junction ŽFig. 4A,B.. An average of 25 neurons double labeled for F-G and DbH ŽFig. 4C,D. were in the most rostral and caudal extents of the DMV. None of the F-G and PNMT double labeled neurons was observed in the DMV following abdominal vagal injection of F-G. Therefore, the great majority of CA containing neurons within the DMV that

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

113

Fig. 3. The photomicrographs illustrate a comparison of F-G labeling in the cranial cervical ganglia and the cervicothoracic ganglia in normal and control injections. A and B are from two different animals receiving F-G injections into both subdiaphragmatic vagal nerve trunks. There are a few F-G labeled neurons with strong fluorescence. C and D illustrate labeling in the same ganglia after 2 ml of F-G were dropped on each vagal nerve trunk rather than injected into the nerve trunk. Fewer neurons are labeled and the fluorescence is very faint. E and F show the SCG and CTG from a rat in which 15 ml of F-G was injected into the abdominal cavity. There are many F-G labeled neurons. The calibration bar is equal to 200 mm.

project to the subdiaphragmatic vagus are dopaminergic Ž88%. and a minority Ž12%. are noradrenergic. Immunopositive reaction for PNMT was absent in F-G labeled neurons in the DMV following abdominal vagal injection of F-G. Neurons, immunostained for TH but not for DbH nor PNMT, were seen in the NG of rats. The mean number of TH immunopositive neurons was 157 Žrange, 124–178. per NG. Of this average, 47 TH neurons were retrogradely labeled by F-G following subdiaphragmatic vagus nerve injection ŽFig. 4E,F.. A number of small TH immunopositive neurons with strong fluorescence were either clustered at the root of the cranial laryngeal branch or scattered among large sized TH immunopositive neurons in the NG.

Small bundles of TH immunopositive nerve fibers were observed passing from the cranial portion of the NG through the extent of the NG, and into the cervical vagus nerve. In addition, we observed several paraganglia with extremely strong TH fluorescence distributed along the cervical and subdiaphragmatic segments Ždid not observe the thoracic segment. of the vagus nerve. The total number of catecholaminergic neurons labeled by F-G from subdiaphragmatic vagal injection is on average equal to 285 neurons. Ninety two percent of this total is of parasympathetic origin with 75% of the total number of neurons lying in the DMV and 17% lying in the NG. The remaining 8% is of sympathetic origin from the cervical and thoracic paravertebral ganglia ŽTable 2..

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

114

Fig. 4. The photomicrographs illustrate the labeling obtained in the DMV and NG with F-G as a retrograde tracer injected into the subdiaphragmatic vagal nerve trunks and immunocytochemistry for the enzymes TH or DBH. Arrows in A and B point to several neurons double labeled for TH and F-G in the caudal DMV; and arrows in C and D point to neurons double labeled for DbH and F-G in the rostral DMV. Arrows in E and F point to several neurons double labeled for TH and F-G in the NG. The calibration bar is equal to 100 mm for A to D and appears in D; the calibration bar is equal to 200 mm for E and F and appears in F.

On either longitudinal or transverse tissue sections of the subdiaphragmatic vagus nerve in rats with F-G injec-

tions, the TH immunopositive nerve fibers in the subdiaphragmatic vagus nerve were of higher density than the

Table 2 Autonomic sources of catecholaminergic ŽCA. neurons projecting to the subdiaphragmatic vagus nerve CA enzymes and F-G double labeling

Sympathetic ganglia Ž n s 5.

Parasympathetic ganglia structures DMV NA NG Ž n s 5. Ž n s 5. Ž n s 8.

Total number of TH and FG neurons in the sympathetic and parasympathetic structures projecting to the subdiaphragmatic vagus nerve

THrF-G Neurons Percentage of total number CA neurons in the autonomic sites

23 8

215 75

285

0 0

47 17

n: Number of rats. Sympathetic ganglia: SCG, CTG and TSG. Parasympathetic ganglia Žparasymp... TH and F-G double labeled neurons: TH is a marker for catecholaminergic neurons; F-G indicates a projection to subdiaphragmatic vagus nerve.

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

115

Fig. 5. The photomicrographs illustrate the labeling obtained in the subdiaphragmatic vagus nerve for TH or DbH labeling. A illustrates nerve fibers labeled for TH distributed in the subdiaphragmatic vagus nerve and B illustrates nerve fibers labeled for DbH from an adjacent section in the same nerve trunk. NV—vagus nerve. The calibration bars are equal to 200 mm and appear on the right of the pair.

DbH immunopositive nerve fibers ŽFig. 5A,B.. The number of TH immunopositive nerve fibers in the abdominal vagus nerve varied in different animals. No PNMT immunopositive nerve fibers were found in the subdiaphragmatic vagus nerve.

4. Discussion 4.1. Methodological considerations An independent test of the success of vagus nerve injections was available through the examination of the labeling found in the NG and the medullary DMV in each animal. The injections generally resulted in labeling a large number of neurons in the NG, confirming an adequate injection procedure. In addition, we used CT-HRP to ensure no peculiar lack of affinity by F-G for this system since so few labeled neurons were found in the sympathetic ganglia after the vagus nerve injection with F-G. In the control experiments, dripping F-G onto or underneath the vagal trunks produced considerably fewer and more lightly labeled neurons in the sympathetic ganglia ŽFig. 3C,D. than direct injections into the vagus nerve ŽFig. 3A,B., but dripping CT-HRP onto the vagal trunks did not result in any significant labeling. We interpret this reduced and faint labeling to be due to F-G spreading from the vagal trunks to the abdomen rather than F-G diffusing into the trunks. This type of labeling is increased in the second control experiment where a larger volume of F-G was injected into the abdomen. Following the injection of F-G into the abdomen to mimic heavy spillage of the tracer, a heavy background of faintly fluorescent labeled neurons

ŽFig. 3E,F. occurred, which we attribute to non-specific spread of F-G through the blood ŽPowley and Berthoud, 1991.. This faint background labeling was never observed in our direct injections, suggesting low leakage of F-G. The average of our stereological estimates of the total number of neurons in the sympathetic ganglia taken from five rats was 36,706 in the SCG and 36,016 in the CTG, which fell within the range of previous estimates of 13,000 to 45,000 ŽSmolen et al., 1983.. The range in total neuron number for these five rats was variable as described in our results. Moreover, we did find that the volume Ž0.52–1.1 mm3 . and neuron density Ž35–63 neuronsrmm3 . of the individual sympathetic ganglion varied considerably from case to case. 4.2. The number and percentage of the catecholaminergic neurons projecting to the subdiaphragmatic Õagus nerÕe The number of F-G single labeled neurons in the sympathetic ganglia varied greatly and was much smaller in comparison with those in the NG and DMV. For example, a consistent low percentage of labeled sympathetic postganglionic neurons to total neuron number was found in the SCG Ž0.016%. and in the CTG Ž0.053%.. Similar results were obtained using CT-HRP. The results suggest that only an extremely small proportion of neurons in the sympathetic ganglia ŽFig. 1B–D, Fig. 2B–D, Fig. 3A–B. travel in the subdiaphragmatic vagus nerve in the rat. The distribution of F-G retrogradely labeled TH immunopositive neurons in the DMV or NG is consistent with previous observations ŽSawchenko et al., 1987; Tayo and Williams, 1988; Ruggiero et al., 1993.. The mean

116

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

number of 215 cells in the DMV was somewhat lower than that reported by Tayo and Williams Ž1988. who found about 335 neurons containing TH. This may be attributable to crushing the subdiaphragmatic vagus nerve before injecting F-G into it. Helke and Rabchevsky Ž1991. reported a decrease in TH neurons of the nodose ganglion as a result of axotomy of the vagus nerve. Our observation that approximately 25 F-G retrogradely labeled DbH immunofluorescent neurons situated mainly at the most rostral and caudal level of the DMV is identical with the observations of Ritchie et al. Ž1982. and Gwyn et al. Ž1985.. As previously reported ŽSawchenko et al., 1987; Tayo and Williams, 1988., we did not see PNMT immunopositive neurons in the DMV. However, Kalia et al. Ž1984. and Ruggiero et al. Ž1993. reported that a few PNMT immunopositive motoneurons exist in the DMV. The discrepancy may be attributable to some PNMT neurons in the DMV innervating organs of the neck or thorax via the vagus nerve because these authors either injected tracer into the nodose ganglion or did cervical vagotomy, which affects more motor fibers than at the subdiaphragmatic level. Since approximately 88% of the F-G and TH immunopositive motor neurons were DbH negative, the parasympathetic catecholaminergic efferents are mainly dopaminergic. This is consistent with previous findings that the concentration of dopamine in the vagus nerve was higher than that of noradrenaline ŽLackovic et al., 1981.. The comparison of the ratio and the nature of the F-G retrogradely labeled catecholaminergic neurons in parasympathetic or sympathetic sources indicate that the majority of the catecholaminergic fibers in the subdiaphragmatic nerve originate from parasympathetic sources Ž92%. and are predominantly dopaminergic in nature, while a very low percentage Ž8%. of the catecholaminergic fibers come from sympathetic sources and are noradrenergic in nature. It is intriguing that approximately 30% of the F-G retrogradely labeled neurons in the cervicothoracic sympathetic ganglia projecting into the vagus nerve were not labeled with TH or DbH. Previous investigators have reported a subpopulation of neurons in sympathetic ganglia of cats and rats that are non-catecholaminergic ŽVerhofstad et al., 1981; Lindh et al., 1989.. These neurons contain acetylcholine, calcitonin gene-related peptide ŽCGRP., vasoactive intestinal peptide ŽVIP. ŽLindh et al., 1989. or serotonin ŽVerhofstad et al., 1981.. It is likely that those non-catecholaminergic neurons in the sympathetic ganglia observed in the present study may contain the above substances.

velopment. Two points merit consideration in this study: Ž1. there is great individual variation in the number of sympathetic postganglionic neurons in the upper cervicothoracic ganglia, e.g. no F-G labeled neurons was seen in 28% of the SCG cases and 14% of the CTG cases; Ž2. the number of sympathetic postganglionic neurons projecting into the abdominal vagus nerve from these sympathetic ganglia is exceedingly small with an average of 0.016% or 0.053% of all neurons in the SCG or CTG, respectively. Hence, these fibers possibly represent an aberration of the autonomic nervous system development. Furthermore, after injecting HRP into the stomach, Hudson Ž1989. found a large number of HRP labeled neurons Ž4172 cells. in the celiac ganglion, but very few labeled neurons in the middle cervical ganglion Ž2 cells. and the sympathetic trunk ganglia Ž32 cells.. He concluded that the paravertebral labeling was of relatively minor importance for control of the stomach. However, these vagal sympathetic neurons could represent sympathetic innervation of blood vessels associated with the vagal trunks. Clearly, the nature and the functional significance of the non-catecholaminergic sympathetic postganglionic nerve fibers remain to be determined. With either the catecholaminergic or non-catecholaminergic fibers of the subdiaphragmatic nerve, their number is so low that their loss, most likely, impacts very slightly in vagotomy experiments where investigators assume a pure parasympathetic effect of vagotomy. The second issue is in regard to the parasympathetic catecholaminergic fibers of the vagus nerve. Consistent with the dopaminergic nature of most of the catecholaminergic vagal preganglionic fibers is the observation that dopamine plays an important role in the protection of gastrointestinal mucosa ŽSzabo et al., 1982; Hernandez et al., 1984, 1987; Glavin, 1989, 1995; Glavin and Hall, 1995; Glavin et al., 1991; MacNaughton and Wallace, 1989.. The latter neurophysiological and neuropharmacological studies showed that dopamine and its agonists attenuate the extent of experimental gut ulcer induced by medications, ethanol, and stress. Dopamine antagonists worsen gastric ulcers by blocking either peripheral DA1 or DA3 receptors in gastrointestinal mucosa ŽGlavin, 1995; Glavin and Hall, 1995.. It is also known that gastric ulcers are a sequelae of Parkinson’s disease ŽGlavin et al., 1991.. We expect that Parkinson’s patients would have a loss of DMV gastric, dopamine containing motoneurons which would remove the protective effect of dopamine from the gastric mucosa and thus account for the gastric ulcers.

Acknowledgements 4.3. Functional considerations The question remains whether this small number of sympathetic postganglionic fibers within the vagus nerve has a special visceral function Že.g. vasomotor. or merely represents an aberration of autonomic nervous system de-

We are grateful to Dr Sheng Chen for his comments on the manuscript and Dr Joel Levine who made the cholera toxin horseradish peroxidase. Yuanyang Liu provided technical assistance. The research was supported by NIH grant GM-27739 awarded to Dr Richard R. Miselis.

M. Yang et al.r Journal of the Autonomic NerÕous System 76 (1999) 108–117

References Ahlman, B.H.J., Larson, G.M., Bombeck, C.T., Nyhus, L.M., 1979. Origin of the adrenergic nerve fibers in the subdiaphragmatic vagus in the dog. Am. J. Surg. 137, 116–122. Coggeshall, R.E., 1992. A consideration of neural counting methods. TINS 15, 9–14. Coggeshall, R.E., Lekan, H.A., 1996. Methods for determining numbers of cells and synapses: a case for more uniform standards of review. J. Comp. Neurol. 364, 6–15. Glavin, G.B., 1989. Activity of selective dopamine DA1 and DA2 agonists and antagonists on experimental gastric lesions and gastric acid secretion. J. Pharmacol. Exp. Ther. 251, 726–730. Glavin, G.B., 1995. A dopamine D3 receptor agonist, 7-hydroxy-N, N-din-propyl-2-aminotetralin, reduces gastric acid and pepsin secretion and experimental gastric mucosal injury in rats. Life Sci. 56, 287–293. Glavin, G.B., Hall, A.M., 1995. Central and peripheral dopamine D1rDA1 receptor modulation of gastric secretion and experimental gastric mucosal injury. Gen. Pharmacol. 26, 1277–1279. Glavin, G.B., Murison, R., Overmier, J.B., Pare, W.P., Bakke, H.K., Henke, P.G., Hernandez, D.E., 1991. The neurobiology of stress ulcers. Brain Res. Rev. 16, 301–343. Gwyn, D.G., Ritchie, T.C., Coulter, J.D., 1985. The central distribution of vagal catecholaminergic neurons which project into the abdomen in the rat. Brain Res. 328, 139–144. Helke, C.J., Niederer, A.J., 1990. Studies on the coexistence of substance P with other putative transmitters in the nodose and petrosal ganglia. Synapse 5, 144–151. Helke, C.J., Rabchevsky, A., 1991. Axotomy alters putative neurotransmitters in visceral sensory neurons of the nodose and petrosal ganglia. Brain Res. 551, 44–51. Hernandez, D.E., Adcock, J.W., Orlando, R.C., Patrick, K.S., Nemeroff, C.B., Prange, A.J. Jr., 1984. Prevention of stress-induced gastric ulcers by dopamine agonists in the rat. Life Sci. 35, 2454–2458. Hernandez, D.E., Mason, G.A., Walker, C.H., Valenzuela, J.E., 1987. Dopamine receptors in human gastrointestinal mucosa. Life Sci. 41, 2717–2723. Hudson, L.C., 1989. The location of extrinsic efferent and afferent nerve cell bodies of the normal canine stomach. J. Auton. Nerv. Syst. 28, 1–4. Kalia, M., Fuxe, K., Goldstein, M., Harfstrand, A., Agnati, L.F., Coyle, J.T., 1984. Evidence for the existence of putative dopamine-, adrenaline- and noradrenaline-containing vagal motor neurons in the brainstem of the rat. Neurosci. Lett. 50, 57–62. Katz, D.M., Markey, K.A., Goldstein, M., Black, I.B., 1983. Expression of catecholaminergic characteristics by primary sensory neurons in the normal adult rat in vivo. Proc. Natl. Acad. Sci. USA 80, 3526– 3530. Katz, D.M., Adler, J.E., Black, I.B., 1987. Catecholaminergic primary sensory neurons: autonomic targets and mechanisms of transmitter regulation. Fed. Proc. 46, 24–29. Kummer, W., Bachmann, S., Neuhuber, W.L., Hanze, J., Lang, R.E., ¨ 1993. Tyrosine-hydroxylase-containing vagal afferent neurons in the rat nodose ganglion are independent from neuropeptide-Y-containing populations and project to esophagus and stomach. Cell Tiss. Res. 271, 135–144. Lackovic, Z., Kleinman, J., Karoum, F., Neff, N.H., 1981. Dopamine and its metabolites in human peripheral nerves: is there a widely distributed system of peripheral dopaminergic nerves? Life Sci. 29, 917–922. Liedberg, G., Nielsen, K.C., Owman, Ch., Sjoberg, N.-O., 1973. Adrener¨ gic contribution to the abdominal vagus nerves in the cat. Scand. J. Gastroenterol. 8, 177–180.

117

Lindh, B., Lundberg, J.M., Holkfelt, T., 1989. NPY-, galanin-, VIPrPHI-, ¨ cGRP- and substance P-immunoreactive neuronal subpopulations in cat autonomic and sensory ganglia and their projections. Cell Tiss. Res. 256, 259–273. Lundberg, J., Ahlman, H., Dahlstrom, ¨ A., Kewenter, J., 1976. Catecholamine-containing nerve fibers in the human abdominal vagus. Gastroenterology 70, 472–474. MacNaughton, W.K., Wallace, J.L., 1989. A role for dopamine as an endogenous protective factor in the rat stomach. Gastroenterology 96, 972–980. Mesulam, M.-M., 1982. Principles of horseradish peroxidase neurochemistry and their applications for tracing neural pathway-axonal transport, enzyme histochemistry and light microscopic analysis. In: Mesulam, M.-M. ŽEd.., Tracing Neural Connections With Horseradish Peroxidase. New York, Wiley, pp. 1–151. Muryobayashi, T., Mori, J., Fujiwara, M., Shimamoto, K., 1968. Fluorescence histochemical demonstration of adrenergic nerve fibers in the vagus nerve of cats and dogs. Jpn. J. Pharmacol. 18, 285–293. Nielsen, K.C., Owman, Ch., Santini, M., 1969. Anatomosing adrenergic nerves from the sympathetic trunk to the vagus at the cervical level in the cat. Brain Res. 12, 1–9. ¨ Ohsumi, K., Tsunekawa, K., Fujiwara, M., 1974. Fluorescence histochemical studies on adrenergic nerve fibers in the vagus nerve of rat. In: Fujiwara, M., Tanaka, C. ŽEds.., Amine Fluorescence Histochemistry. Tokyo, Igaku Shoin, pp. 93–104. Pakkenberg, B., Gundersen, H.J.G., 1988. Total number of neurons and glial cells in human brain nuclei estimated by the disector and the fractionator. J. Microsc. 150, 1–20. Powley, T.L., Berthoud, H.-R., 1991. A fluorescent labeling strategy for staining the enteric nervous system. J. Neurosci. 36, 9–15. Ritchie, T.C., Westlund, K.N., Bowker, R.M., Coulter, J.D., Leonard, R.B., 1982. The relationship of the medullary catecholamine containing neurons to the vagal motor nuclei. Neuroscience 7, 1471–1482. Ruggiero, D.A., Chau, L., Anwar, M., Mtui, E.P., Golanov, E.V., 1993. Effect of cervical vagotomy on catecholminergic neurons in the cranial division of the parasympathetic nervous system. Brain Res. 617, 17–27. Sawchenko, P.E., Cunningham, E.T., Jr., Levin, M.C., 1987. Anatomic and biochemical specificity in central autonomic pathways. In: Ciriello, J., Calaresu, F.R., Renaud, L.P., Polosa, C. ŽEds.., Organization of the Autonomic Nervous System: Central and Peripheral Mechanisms. Alan R. Liss, New York, pp. 267–281. Smolen, A.J., Wright, L.L. unnin, Cunningham, T.J., 1983. Neuron numbers in the superior cervical sympathetic ganglion of the rat: a critical comparison of methods for cell counting. J. Neurocytol. 12, 739–750. Szabo, S., Sandrock, A.W., Nafradi, J., Maull, E.A., Gallagher, G.T., Blyzniuk, A., 1982. Dopamine and dopamine receptors in the gut: their possible role in duodenal ulceration. Adv. Biosci. 37, 165–170. Tayo, E.K., Williams, R.G., 1988. Catecholaminergic parasympathetic efferents within the dorsal motor nucleus of the vagus in the rat: a quantitative analysis. Neurosci. Lett. 90, 1–5. Verhofstad, A.A.J., Steinbusch, H.W.M., Penke, B., Varga, J., Joosten, H.W.J., 1981. Serotonin-immunoreactive cells in the superior cervical ganglion of the rat. Evidence for the existence of separate serotoninand catecholamine-containing small ganglionic cells. Brain Res. 212, 39–49. West, M.J., Slomianka, L., Gundersen, H.J.G., 1991. Unbiased stereological estimation of the total number of neurons in the subdivisions of the rat hippocampus using the optical fractionator. Anat. Rec. 231, 482–497.