Autonomic innervation of the human middle ear: An immunohistochemical study

Original Contributions Autonomic Innervation of the Human Ear: An Immunohistochemical Study B. Shiva Nagaraj, MD, and Fred H. Linthicum,

Middle

Jr, MD

Purpose: Although there have been numerous studies of autonomic innervation of the middle ear mucosa, and the mechanism of effusion into the middle ear cavity in animals, the autonomic innervation of the human middle ear has not received much attention. The purpose of this study is to show the presence of catecholaminergic nerve fibers in the human middle-ear mucus membrane that may play an important role in the pathogenesis of middle-ear effusion. Materials and Methods: A total of 126 celloidin-embedded temporal bone sections from the temporal bone bank at the House Ear Institute were used for immunohistochemical study. A polyclonal antibody to tyrosine hydroxylase enzyme was used to show the presence of catecholaminergic nerve fibers. Results: Tyrosine hydroxylase immunoreactive nerve fibers containing numerous fine varicosities along their course, characteristic of noradrenergic neurons, were observed throughout the middle-ear mucosa including the promontary, sinus tympani, mesotympanum, and hypotympanum. In addition, these nerve fibers were seen in close promixity to the small-caliber blood vessels. A striking variation in the intensity of staining as well as in the amount of nerve fibers was observed among the temporal bone sections. Conclusion: It is possible that the catecholaminergic nerve fibers, like elsewhere in the body, may exert a direct influence on the middle-ear mucosal blood vessels. We speculate that the effusion into the middle-ear space is an active, rather than a passive process. It is conceivable that cholinergic-sympathetic nerves might exist in the human middle-ear mucus membrane, and that these autonomic nerves, in conjunction with the neuropeptides, may play an active role in the pathogenesis of human middle-ear effusion. Copyright 0 1998 by W.B. Saunders Company.

Immunocytochemical studies have shown the presence of nerve fibers containing neuropeptides such as vasoactive intestinal peptide (VIP), substance P (SP), and immunoreactive avian pancreatic polypeptide (APP) in the nasal mucosa and tracheobronchial wall.*-3 The distribution of nerve fibers containing neuropeptides, norepinephrine (NE), and acetylcholine (ACh) in the mammalian middleear mucosa (MEM), as well as the tympanic membrane (TM), has been rather extensively

From the Temporal House Ear Institute,

Bone Histopathology Los Angeles, CA.

Laboratory,

Address reprint requests to Fred H. Linthicum. Director, House

Temporal Ear Institute.

Bone 2100

Histopathology W Third St.

Jr, MD,

Laboratory; 5th Floor. Los

Angeles, CA 90057-9927. Copyright 0 1998 by W.B. Saunders 0196-0709/98/l 902-0001$8.00/O American

Journal

Company

of Otolatyngology,

studied using immunocytochemical and histofluorescence techniques4J Although the animal studies are numerous, the autonomicinnervation of the human middle ear (ME) has received less attention. Animal studies propose the attic space to be the initial site of effusion production into the ME cavity.6,7 Autonomic nerve fibers exhibiting SP-like, VIPlike immunoreactivity, have been shown to be localized exclusively to the pars flaccida of the mammalian ear drum, and are thought to play an important functional role in the pathogenesis of otitis media with effusion (OME) in experimental animals.8 In the present study, we have analyzed the MEM of human temporal bone sections using immunohistochemistry to show catecholaminergic nerve fibers. The presence of tyrosine hydroxylase (TH) immunoreactive nerve Vol 19, No 2 (March-April),

1998:

pp 75-82

75

76

NAGARAJ

fibers in the MEM would suggest that the sympathetic nerves, like elsewhere in the body, may exert a direct influence on the mucosal blood vessels. There is a possibility for the colocalization of TH and ACh in the autonomic nerves innervating the MEM. We speculate that cholinergic-sympathetic nerves, which are anatomically sympathetic but are functionally cholinergic, similar to the sympathetic nerves innervating the human sweat glands, might be involved in the secretory functions of the MEM. The possible presence of cholinergic-sympathetic nerve fibers in the MEM

may

be one

of the

mechanisms

in the

pathogenesis of effusion into the ME. Further research is needed to substantiate the presence of cholinergic-sympathetic nerves in the MEM

and

their

possible

putative

role

in the

pathogenesis of middle-ear effusion (MEE). Antibodies to various catecholamine synthesizing enzymes have been developed to identify catecholamine-containing neurons. Anti-TH antibody in immunostaining

tive

sympathetic

has been extensively used studies to show TH-posi-

(catecholaminergic)

nerves

in various animal and human tissues. In our study, a polyclonal antibody against TH, antiTH, was used to identify the presence of catecholaminergic neurons within the MEM of

human temporal

MATERIALS

bone sections.

AND METHODS

Tissue Specimen The human temporal bone specimens used in our study were selected from the House Ear Institute’s temporal bone bank. Temporal bone specimens from individuals with a clinical history of ME disease, and the specimens showing postmortem degradation changes, were excluded from this study. These bone specimens had been processed routinely by fixation with 10% neutral buffered formalin and subsequent decalcification with ethylenediamine-tetra-acetate (EDTA). Using a sliding microtome, bone sections measuring about 20 pm in thickness were sliced from the selected specimens. Every 10th section was stained with hematoxylin and eosin and mounted on a microslide for histopathological testing. The remaining temporal bone sections were stored in glass jars containing 80% ethyl alcohol, and were available for immunohistochemical study. A total of 126 celloidinembedded bone sections from 10 different individuals between 55 and 85 years of age were used in the present study.

AND

LINTHICUM

Antisera Primary antibody. Rabbit anti-TH polyclonal antibody came from Chemicon International Inc, Temecula, CA. Final concentration of the primary antibody used for obtaining positive staining was 1:1,500.

Secondary antibody. Biotinylated anti-rabbit immunoglobulin came from Biogenex laboratories, San Ramon, CA. Final concentration of the secondary antibody used for obtaining positive staining was 1:20.

lmmunohistochemical

Staining

Procedures

The immunostaining technique was systematically varied several times to develop the correct protocol and to establish optimal working dilutions in order to obtain positive staining. Day one. The temporal bone sections were dehydrated by successively immersing them, for 10 minutes each, in 80%, 95%, 100% ethyl alcohol, and finally in 100% methanol solutions. Celloidin was removed from the bone sections, by treating with NaOH-methanol solution (dilution 1:4) for 15 minutes, to retrieve antigens masked by formalin fixation and to improve the immunoreactivity of the epitopes. Tissue-sections were then washed consecutively in 100% and in 80% methanol for 10 minutes each, and then rehydrated with phosphatebuffered saline (PBS) at pH 7.4. The sections were cleaned with 0.03% Triton X-100 (Research Organits Inc, Cleveland, OH) for 10 minutes to increase the permeability and then rinsed in PBS for 10 minutes. Next, full strength ficin solution was added to the sections to unmask the antigenic sites hidden by protein cross links. The sections were covered with parafilm and incubated in the oven at 60°C for 15 minutes. Sections were again washed in two changes of PBS for 10 minutes each, and then blocked with 1% bovine serum albumin-PBS for 30 minutes. Anti-TH in 1:1,500 dilution was applied to the sections, covered with parafilm, and incubated overnight at 4°C. Day two. The sections were washed in two changes of 0.5% casein-PBS for a total of 40 minutes. Biotinylated secondary antibody (rabbit-link antibody) in 120 dilution was applied to the sections, covered with parafilm, and incubated for 1 hour at room temperature. The sections were again washed in two changes of 0.5% casein-PBS for a total of 40 minutes. In the next step, the sections were incubated with Vectastain Elite ABC kit (Avidin DH and biotinylated horseradish peroxidase H macro molecular complex; Vector Laboratories, Burlingame, CA) label for 1 hour at room temperature. After this, the sections were stained with diaminobenzidine tetrahydrochloride. Sections were mounted on glass slides in PBS, and air dried overnight.

AUTONOMIC

INNERVATION

OF THE

HUMAN

MIDDLE

EAR

Day three. Finally, the sections were dehydrated through 80%, 95%, loo%, 100% ethyl alcohol, cleared with xylene, and coverslipped using permount. In this study PBS was used in the negative control specimens.

RESULTS The expression

of TH activity,

an adrenergic

marker, was observed in the ME mucus membrane. On light microscopic examination, nerve fibers containing numerous fine varicosities along their course, characteristic of noradrenergic neurons, were seen throughout the mucus membrane of the ME (Figs 1 and Z), including the promontary, sinus tympani, mesotympanum (Fig 3), and the mucosa of the air cells in the hypotympanum (Fig 4). In addition, nerve fibers exhibiting TH immunoreactivity were seen in close proximity to smallcaliber blood vessels in the MEM (Figs 3 and 4), and in the mucosa of mastoid air cells. In contrast to animal studies, we were not able to show TH-positive nerve fibers in the TM. Also, no TH-positive nerve fibers were found in the endolymphatic sac, vestibular organs, cochlea, or the eustachian tube. A striking variation was observed both in the intensity of staining and in the amount of nerve fibers among the temporal bone sections from one temporal bone specimen to the next as well as within the same individual bone specimens. In the negative control temporal bone sections, no staining of nerve fibers could be detected.

Fig 1. Temporal bone section showing TH-positive nerve fibers containing numerous fine varicosities along their course (arrow) in the ME mucous membrane in the proximity of sinus tympani. (Bar = 0.2 mm.)

77

DISCUSSION NE, epinephrine, and dopamine are the principal catecholamines found in the body, and are important chemical transmitters. TH enzyme plays an important role in the catecholamine biosynthesis. A brief review of catecholamine physiology at this juncture is useful in order to appreciate the rationale of both the present study and future research in this field. Catecholamines are formed by hydroxylation and decarboxylation of the aminoacids phenylalanine and tyrosine, and several important enzymes are involved in their biosynthesis (Fig 5). Phenylalanine hydroxylase, which is found primarily in the liver, catalyzes the conversion of phenylalanine to tyrosine. Tyrosine is then transported into catecholaminesecreting neurons by a concentrating mechanism and is converted to dihydroxyphenylalanine (DOPA). This step involving the conversion of tyrosine to DOPA is catalyzed by TH to dopamine enzyme. DOPA is then converted by dopa decarboxylase. DOPA and dopamine are formed in the neuronal cytoplasm. Dopamine enters the granulated vesicles, within which it is converted to NE by dopamine p-hydroxylase. In the adrenal medulla, the granules in some of the cells contain the enzyme phenylethanolamine-N-methyltransferase, which catalyzes the conversion of NE to epinephrine. The rate-limiting step in the biosynthesis of catecholamines is the conversion of tyrosine to DOPA. TH, which catalyses

78

NAGARAJ

AND

LINTHICUM

Fig 2. Temporal bone section showing catecholaminergic nerve fibers with beaded appearance (arrow) in the ME mucous membrane. (Bar = 0.2 mm.)

this key step, is in turn subject to feedback inhibition by dopamine and NE, thus providing internal control of the synthetic process. Catecholamines are released from autonomic neurons and adrenal medullary cells by exocytosis, along with adenosine triphosphate, dopamine P-hydroxylase, chromogranin A, and related proteins. The dopaminesecreting, NE-secreting, and epinephrinesecreting neurons are appropriately called dopaminergic, noradrenergic, and adrenergic neurons. NE, also known as noradrenaline or levarterenol. is the chemical transmitter at most

Fig 3. Nerve fibers exhibiting tyrosine hydroxylase immunoreactivity are seen as black lines with varicosities close to the blood vessels (arrow) in the mucosa of the mesotympanum. anterior (Bar = 0.25 mm).

sympathetic postganglionic endings. It is stored in the synaptic knobs of neurons, which secrete it in granulated vesicles that have a dense core. Generally, the sympathetic postganglionic neurons have characteristic multiple varicosities along their course, and each of these varicosities appears to be a site at which NE is liberated. NE and its methyl derivative, epinephrine, are secreted by the adrenal medulla, but epinephrine is not a mediator at the sympathetic postganglionic endings. Transmission at the synaptic junctions between preganglionic and postganglionic neurons, and between the postganglionic neurons

AUTONOMIC

INNERVATION

OF THE

HUMAN

MIDDLE

EAR

Fig 4. (A) Catecholaminergic nerve fibers appearing as black lines are seen in close proximity to the blood vessels (arrow) in the hypotympanic cell wall. (Bar = 0.25 mm). (B) Photomicrograph of the negative control specimen showing absence of staining of the nerve fibers adjacent to the blood vessels.

and the autonomic effector organs, is chemically mediated. The principal chemical transmitters in the autonomic nervous system are ACh and NE. In addition, dopamine is secreted by interneurons (dopaminergic) in the sympathetic ganglia. On the basis of the chemical transmitter released, the autonomic nervous system can be broadly divided into two chemical divisions, cholinergic (ACh-secreting), or parasympathetic, and noradrenergic (NE-secreting), or sympathetic divisions. Besides all the preganglionic neurons and the anatomically parasympathetic postganglionic neurons that

are cholinergic, the anatomically sympathetic post-ganglionic neurons that innervate sweat glands and the anatomically sympathetic neurons that end on blood vessels in skeletal muscles (sympathetic vasodilator nerves) are also cholinergic. The remaining sympathetic postganglionic neurons are noradrenergic.g The TM of all laboratory animals and humans is composed of three distinctive layers: an outer epidermal layer, a middle lamina propria, and an inner mucosal epithelium. The TM typically consists of two parts, pars flaccida (Shrapnell’s membrane) and pars tensa. The pars flaccida contains a well-

80

NAGARAJ

NH,

HLCOOH

d-COO” I

1 Phenylalmnc hydroxyla‘e

Tyrosine hydroxylrse

-0 I

HO-

bH Phenylalrnme

,’

‘\

I NH,

Al,

+n2

“L-OH

HC-OH

DOM IDihvdroxvphhenvlal~nine)

bH

I

DOW demboxylase

\ \

/

NH2

\

I

LHI ’ W

\

Dopamine f3hydroxylue HO OH Epinephrina

/

’

Tyrorine /

HN-CH,

LINTHICUM

NH,

NH,

WC-COOH

AND

0

HO-’ OH Norepinsphrine

developed lamina propria and, contrary to common belief, is thicker than the pars tensa. The pars tensa, on the other hand, is characterized by its unique fibrous layer that provides tension to the TM. The fibrous layer consists of an outer radial and an inner circular layer of collagenous fibrils made up of types II and III collagens to a large extent, and type 1 collagen to a lesser extent.‘O Besides sound transmission, the TM plays an important role in the ME defense mechanism. Specialized encapsulated nerve endings, that have been regarded as “modified” pacinian corpuscles, have been found in the TM,” and the TM is thought to function as a baroreceptor capable of detecting rapidly fluctuating pressure changes in the ME. The TM also has a rich vascular supply, a rich network of nerves containing neuropeptides, and a rich storage of mast cells, that collectively participate in the neurogenic inflammation of the ME. Under normal conditions, the TM in a healthy ear has minimal blood vessels. In inflammatory conditions of the ME cavity, the appearance of the TM is significantly altered. Thus, in acute otitis media, the pars flaccida portion of the TM shows diffuse vascular congestion in the early stage, whereas in otitis media with effusion, dilated radiating vessels are seen around the periphery of the pars tensa. Stimulation of sensory nerves supplying the TM elicits vasodilatation and an increase in vascular permeability. This response, termed neurogenic inflammation, is

‘\ /

Fig 5. Biosynthesis of catecholamines. The dashed lines indicate inhibition of TH by NE and dopamine. Tetrahydrobiopterin is a cofactor for the action of phenyalanine hydroxylase and TH. (Reprinted with permission.)g

Dopmim

IDihydroxyphenyl~ ethylaminsl

OH

mediated by sensory neuropeptides such as SP, VIP, calcitonin gene-related peptide (CGBP), and neurotransmitters such as ACh.12 The pathogenesis of OME has been the subject of much speculation for many years. In addition to many factors such as eustachian tube obstruction, infection by viruses and bacteria, immunologic factors, and mechanical stimulation of the external auditory cana1,13 neurogenic inflammation has been suggested as an important pathogenic factor in the mechanism of OME production.12 Sensory nerve fibers exhibiting SP-like, VIPlike, and enkephalin-like immunoreactivity have been found to be particularly abundant in the pars flaccida of the TM of ratse8 Immunohistochemical studies have revealed SP-like immunoreactivity in the tympanic nerve of humans and various animal species. SP is thought to play a putative role in the neurochemical control of ME aeration, in addition to its role in the neurogenic inflammation of the MEM.14 Hentzer15 has described mast cells in the submucosa of the human MEM, and Widemar et all” have shown that the Shrapnell’s membrane of the human TM has a rich collection of mast cells in its connective tissue layer. The neuropeptide SP has been reported to cause mast cell degranulation with the release of histamine and to evoke marked vascular leakage. 17,18The vascular effects of SP can be attributed to its direct action on the blood vessels, as well as its ability to cause histamine release from the mast cells.12 Likewise, VIP has been shown to provoke vasodila-

AUTONOMIC

INNERVATION

OF THE

HUMAN

MIDDLE

EAR

tation and edema formation probably by the same mechanism as SP.lg Studies on mammalian TM have shown that, whereas both ACh and VIP evoke vasodilatation and vascular leakage into the ME cavity, the alpha adrenergic receptors mediate vasoconstriction of the TM blood vessels.20 The CGRP immunoreactive neurons have been shown to coexist with the neuropeptide SP, and to cause vasodilatation and potentiation of edema.21 In some recent animal studies, the attic space has been proposed to be the initial site of effusion production in OME.“s7 The pars flaccida, rich in mast cells and edema-provoking neuropeptides, shows the earliest visible changes, such as vasodilatation, discoloration, and edema formation when MEE is experimentally provoked in animals.* This finding may be of importance in establishing the pars flaccida portion of the TM as the initial site of effusion production in experimental serous otitis media. The edema and vascular changes in the pars flaccida seem to occur concomitantly with mast cell degranulation and elevated levels of histamine in the ME cavity.22 This observation indicates that the mast cells may be involved in the onset of MEE, and the neural elements may play a direct role in the trigger mechanism of effusion production. The foregoing discussion suggests that the interactions between the histamine-containing mast cells in the MEM and the neuropeptides produce neurogenic inflammation, and this could be one of the mechanisms involved in the pathogenesis of OME. Although TH is a rate-limiting enzyme in catecholamine synthesis, a weak TH and neuropeptide Y immunoreactivity has been described in cholinergic nerves of the iris in sympathectomized rats.23 Studies by Landis et alz4 indicate that the cholinergic-sympathetic neurons appear to undergo a transition from noradrenergic to cholinergic function in vivo. Thus, there is evidence to suggest that the autonomic nerves in certain locations appear to exhibit neurotransmitter plasticity and developmental changes in their properties. In humans, it has been established that the anatomically sympathetic neurons innervating the sweat glands and the blood vessels in skeletal muscles are cholinergic in their function.9 It is a hypothetical possibility that such cholinergic-sympathetic nerves might be present in the

81

MEM, and we speculate that this might be one of the mechanisms in the pathogenesis of MEE. Coexistence of a neurotransmitter with neuropeptides is a well-established phenomenon. In our study, only the immunoreactivity to TH in the MEM was examined. Doublelabel immunostaining studies are needed (1) to confirm the possible coexistence of neuropeptides with the neurotransmitters in the nerves innervating the MEM and the possible colocalization of TH in the cholinergic nerves supplying the MEM and (2) to lend credence to the hypothesis that cholinergic-sympathetic nerves may be involved in the secretory functions of the MEM and in the pathogenesis of MEE. CONCLUSION The present study has shown that the human MEM is supplied with catecholaminergic nerve fibers. It is conceivable that these nerves may exert a direct influence on the mucosal blood vessels. Based on the observations in animal studies, it is apparent that the neuropeptides serve as local regulators of vasomotor function and mediate neurogenic inflammation in the ME. We speculate that the MEE production is an active, rather than a passive, process. It is possible that cholinergic-sympathetic nerves might exist in the human MEM, and these nerves, in conjunction with the neuropeptides, may play a role in the pathogenesis of MEE. Further research in the field of neuropeptides and the autonomic innervation of human ME is needed to fully understand the mechanism of effusion production. This knowledge may help us to develop pharmacological agents and to design treatment methods that could hinder the pathological process from becoming manifest. ACKNOWLEDGMENT The authors are grateful to William H. Slattery III, MD, for valuable suggestions and comments, and to Karen Berliner, PhD, for editing the manuscript. We also thank Diana Cohen for her help in the laboratory and Liz Gnerre for library services. REFERENCES 1. Uddman and distribution

R, Alumets J, Densert 0, et al: Occurrence of VIP nerves in the nasal mucosa and

82

tracheobronchial wall. Acta Otolaryngol (Stockh) 86:443448,1978 2. Uddman R, Malm L, Sundler F: Peptide containing nerves in the nasal mucosa. Rhinology 19:75-79,198l 3. Uddman R, Hakanson R, Sundler F: Immunoreactive avian pancreatic polypeptide occurs in nerves of the mammalian nasal mucosa and eustachian tube. ORL J Otorhinolaryngol Relat Spec 42:242-248,198O 4. Ishii T, Kaga K: Autonomic nervous system of the cat middle ear mucosa. Ann Otol Rhino1 Laryngol Suppl 85:51-57, 1976 (suppl25) 5. Uddman R, Kitajiri M, Sundler F: Autonomic innervation of the middle ear. Ann Otol Rhino1 Laryngol 92:151-154,1983 6. Hellstrom S, Salen B, Stenfors L-E: The site of initial production and transport of effusion materials in otitis media serosa. A study on rat middle ear cavity. Acta Otolaryngol (Stockh) 93:435-440, 1982 7. Stenfors L-E, Hellstrom S, Salen B: The role of the attic space in experimental otitis media with effusion. Acta Otolaryngol (Stockh) 386:146-148, 1982 (suppl) 8. Widemar L, Hellstrom S, Schultzberg M, et al: Autonomic innervation of the tympanic membrane: An immunocytochemical and histofluorescence study. Acta Otolaryngol (Stockh) 100:58-65, 1985 9. Ganong WF (ed): Synaptic and junctional transmission (chap 4) and the autonomic nervous system (chap 13), in Review of Medical Physiology (ed 17). Norwalk, CT, Appleton&Lange, 1995, pp 89-100 and 203-207 10. Lim DJ: Structure and function of the tympanic membrane: A review. Acta Otorhinolaryngol Belg 49:101115,1995 11. Nagai T, Tono T: Encapsulated nerve endings in the human tympanic membrane. Ann Otol Laryngol 246:169172,1989 12. Hellstrom S, Goldie P: Mechanisms of otitis media development-Involvement of neurogenic inflammation. Otolaryngol Clin North Am 24:829-834,199l 13. Alm PE, Bloom GD, Hellstrom S, et al: Middle ear effusion caused by mechanical stimulation of the external auditory canal. An experimental study in the rat. Acta Otolaryngol (Stockh) 96:91-98,1983 14. Ylikoski J, Panula P: Neuropeptides in the middle

NAGARAJ

AND

LINTHICUM

ear mucosa. ORL J Otorhinolaryngol Relat Spec 50:176182,1988 15. Hentzer E: Histologic studies of the normal mucosa in the middle ear, mastoid cavities, and eustachian tube. Ann Otol Rhino1 Laryngol 79:825-833,197O 16. Widemar L, Hellstrom S, Stenfors L-E: An overlooked site of tissue mast cell-The human tympanic membrane. Acta Otolaryngol (Stockh) 102:391-395, 1986 17. Fewtrell CMS, Foreman JC, Jordan CC, et al: The effects of substance P on histamine and 5-hydroxytryptamine release in the rat. J Physiol London 336:393411,1982 18. Hagermark 0, Hokfelt T, Pernow B: Flare and itch by substance P in human skin. Invest Dermatol 71:233235,1978 19. Williams TJ: Vasoactive intestinal polypeptide is more potent than prostaglandin E2 as a vasodilator and edema potentiator in rabbit skin. Br J Pharmacol 77:505509,1982 20. Hellstrom S, Albiin N, Goldie P, et al: Pharmacological characterization of receptors on blood vessels in the tympanic membrane involved in otitis media. Auris Nasus Larynx (Tokyo) 12:S135-S137,1985 (suppl 1) 21. Uddman R, Grunditz T, Larsson A, et al: Sensory innervation of the ear drum and middle ear mucosa: Retrograde tracing and immunocytochemistry. Cell Tissue Res 252:141, 1988 (abstr) 22. Alm PE, Bloom GD, Hellstrom S, et al: The release of histamine from the pars flaccida mast cells. One cause of otitis media with effusion? Acta Otolaryngol (Stockh) 94:517-522,1982 23. Bjorklund H, Hokfelt T, Goldstein M, et al: Appearance of the noradrenergic markers tyrosine hydroxylase and neuropeptide Y in cholinergic nerves of the iris following sympathectomy. J Neurosci 5:1633-1643, 1985 24. Landis SC, Keefe D: Evidence for neurotransmitter plasticity in vivo: Developmental changes in properties of cholinergic sympathetic neurons. Dev Biol 98:349-372, 1983 25. Hokfelt T, Lundberg JM, Skirboll L, et al: Coexistence of classical transmitters and peptides in neurons, in Cue110 AC (ed): Co-transmission. London, UK, MacMillan, 1982, pp 77-126