INTRODUCTION

TOP ABSTRACT INTRODUCTION NONADRENERGIC/NONCHOLINERGIC NONADRENERGIC/NONCHOLINERGIC PROBLEMS REFERENCES

The nonadrenergic/noncholinergic (NANC) nervous control of airway smooth muscle has been extensively studied in the past two or three decades, with a vast and ever-growing literature. However, its importance, at least in human health and disease, remains unestablished and controversial. The inhibitory (i-NANC) system may modulate smooth muscle tone, and the excitatory (e-NANC) system certainly underlies neurogenic inflammation in experimental animals, but both remain to be convincingly demonstrated in humans.

The terms i-NANC and e-NANC are unsatisfactory. They apply only to the nervous actions on airway smooth muscle, but the same released transmitters will have other actions (Table 1). For example, the e-NANC transmitters substance P (SP) and neurokinin A (NKA) will contract airway smooth muscle but will relax that of the mucosal vasculature. Neuropeptide Y (NPY), potentially an i-NANC transmitter but not usually regarded as such, will relax airway smooth muscle but contract that of the vasculature (1). Actions on submucosal glands are complex, and may be more difficult to define in terms of excitation and inhibition. This is more than a semantic point. Insofar as the NANC transmitters change blood flow, they will potentially alter the output of nitric oxide (NO) and endothelins from the vascular endothelium, and these mediators may in turn influence airway smooth muscle tone. The e-NANC tachykinins also cause profound, if transient, changes in epithelial structure, and both of the better-established i-NANC transmitters, NO and vasoactive intestinal polypeptide (VIP), can act on the epithelium (2, 3); these actions would almost certainly affect the release of various mediators derived from the epithelium and acting on airway smooth muscle. Thus, it is unlikely that the individual NANC transmitters, usually studied in isolation and in action of isolated target organs, would in vivo have an independent action. Rather their release would, at least in theory, lead to a chain of actions due to the secondary release of other mediators.

The criteria for establishing that a chemical is a neurotransmitter, as applied to the NANC system, are listed in Table 2. Whereas not all criteria can be applied to all potential neurotransmitters, for example, specific antagonists or potentiating agents may not be available, one would need most of the criteria to be met for a claim to neurotransmission to be convincing. With regard to airway smooth muscle, only three agents meet all or nearly all these criteria: NO, VIP, and the tachykinins SP/NKA. Others, such as NPY, meet some of the criteria, but many agents known to be present in airway autonomic nerves and with the potential to be neurotransmitters have a far more speculative role. These include, for example, galanin, somatostatin, gastrin-relaxing factor, enkephalin, calcitonin gene-related peptide (CGRP), and several others. These will not be further considered here. Even when criteria are met to establish a neurotransmitter, other questions remain to be answered. Do the concentrations shown to be effective in experimental (usually in vitro) situations also apply under more physiologic conditions? Can the mediator release and action be activated by physiologic processes, for example, by reflex (central or local) mechanisms? Can the mechanisms, nearly always investigated in experimental animals, be shown also to be present in humans? If so, are they important or just other minor curiosities similar to the many that litter the literature?

	NONADRENERGIC/NONCHOLINERGIC INHIBITORY (i-NANC) NERVES

TOP ABSTRACT INTRODUCTION NONADRENERGIC/NONCHOLINERGIC NONADRENERGIC/NONCHOLINERGIC PROBLEMS REFERENCES

Vasoactive Intestinal Polypeptide

VIP meets nearly all the criteria of Table 2 (4, 5). For example, it is present in airway nerves and ganglia, as shown by fluorescent microscopy (6), and coexistent with acetylcholinesterase, which suggests that the nerves are motor parasympathetic fibers. The mode of synthesis of VIP has been described. It has been measured on release by nerve stimulation. It is a small diffusible molecule, which applied exogenously relaxes airway smooth muscle, together with several other important motor actions. It can be broken down by local enzymes such as neutral endopeptidase (NEP), and mast cell tryptase and chymase. The effects of exogenous and endogenous VIP can be prevented by -chymotrypsin. The general consensus is that VIP is an important i-NANC neurotransmitter that relaxes airway smooth muscle.

Some problems require more study. VIP, on immunohistologic evidence, is co-localized with several other putative transmitters, both in airway nerves and in ganglia (4, 6); these include especially NO, but also peptide histidine methionine (PHM), peptide histidine isoleucine (PHI), various opioids, galanin, CGRP, and NPY. If these rather nonquantitative histologic studies have physiologic relevance, they must mean that the nervous activity that releases VIP and acetylcholine (ACh) will also release a host of other mediators with various actions on effector tissues.

A second major consideration is that several experimental studies have shown that VIP, although even more abundant in airway nerves in humans than in other species, such as the guinea pig (4), is probably not the main parasympathetic NANC inhibitor in humans (3, 10). Thus, i-NANC relaxation of bronchial smooth muscle in humans is not blocked by the VIP antagonist -chymotrypsin, whereas it is by inducible NO synthase (iNOS) inhibitors (see below). And in humans exogenous VIP has not been established as an effective bronchodilator agent, although this could be because there are barriers in its ability to reach airway smooth muscle.

In summary, all the evidence points to VIP being an important or the main i-NANC smooth muscle relaxant mediator in experimental animals, especially the guinea pig, but its importance in humans is dubious. In which case what is the role of VIP, the abundant presence of which in human airway nerves is well established?

Nitric Oxide

Like VIP, NO fills nearly all the criteria for an i-NANC transmitter (3, 10). iNOS is present in airway nerves, on abundant histologic evidence (11). It would diffuse rapidly to airway smooth muscle. It is inactivated by combination with hemoglobin in the mucosal vasculature. Local application of NO or its donor, nitroprusside, relaxes airway smooth muscle. And NOS inhibitors suppress the action of i-NANC activation in airway preparations with field stimulation, at least in humans (16). Its mechanism of release from nerves is unclear; it is said to be present in solution in the nerve cytoplasm, and to respond to "local demand," but the way this works does not seem to have been established.

The evidence points to NO being an important or the main i-NANC transmitter, at least in humans. In other species the agent is present and may play a collaboratory role with VIP. As with VIP, the co-localization of NO with other transmitters raises questions that need to be answered. The ubiquitous presence of NO in vascular endothelium, in the epithelium of guinea pigs (11) and asthmatic (14) and possibly healthy (13, 15) humans, in sympathetic motor nerves to the lungs, and even in epithelial nerves (11), presumed sensory, emphasizes the importance of NO as a mucosal mediator, but does not particularly help our understanding as to how it plays this role. A further consideration is that in guinea pigs, and probably in humans, the amount of NO in the epithelium greatly exceeds that in airway nerves.

Physiologic Activation of i-NANC

Almost all the studies on the i-NANC have been with in vitro preparations, with field stimulation of nerves or with excitation of the attached vagal (parasympathetic) trunks. It is more difficult to study the natural activation of the mechanism, which must be either by reflexes or from the central nervous system. In experimental animals these reflexes have been extensively studied (17, 18). Among others, bronchoconstrictor reflexes can be elicited on stimulation of laryngeal and tracheobronchial rapidly adapting (RAR, cough) and C-fiber receptors, pulmonary C-fiber receptors, peripheral and central chemoreceptors, and irritant receptors in the nose. Bronchodilator reflexes can be induced by mechanical stimulation of the nose or nasopharynx, stimulation of slowly adapting pulmonary stretch receptors, and from skeletal muscle receptors and baroreceptors. All these reflexes can be blocked by atropine, but this agent leaves the smooth muscle fully relaxed, so any i-NANC inhibition of smooth muscle tone would not be apparent; therefore, for most of these reflexes the possible involvement of the I-NANC system has not been shown.

However, in cats both laryngeal irritation and pulmonary C-fiber stimulation cause a vagally mediated bronchodilatation when cholinergic and adrenergic mechanisms are blocked (by atropine and propranolol), and some degree of bronchomotor tone has been pharmacologically restored (19). In humans a similar bronchodilator response can be observed on mechanical stimulation of the larynx (20). These experiments provide convincing evidence that the motor i-NANC pathway can be reflexively activated in both animals and humans, although the nature of the i-NANC transmitter has not been determined.

	NONADRENERGIC/NONCHOLINERGIC EXCITATORY (e-NANC) NERVES

TOP ABSTRACT INTRODUCTION NONADRENERGIC/NONCHOLINERGIC NONADRENERGIC/NONCHOLINERGIC PROBLEMS REFERENCES

Neurogenic inflammation is the group of local responses that occur when sensory nerves in the mucosa are stimulated to release sensory neuropeptides, such as the tachykinins SP and NKA, and also CGRP (23). The release of SP and NKA has been measured by immunofluorescent methods. Stimulation of the nerves sets up an axon reflex in the distribution of the sensory nerve terminals (24). These neurotransmitters can cause vasodilatation with increased plasma extravasation and exudation into the airway lumen, mucus secretion, smooth muscle contraction, and structural changes in the epithelium. It has been suggested that the term "neurogenic inflammation" should be restricted to the plasma extravasation and exudation, but in terms of mechanisms all the motor actions may be present.

All the criteria in Table 2 seem to be met, at least for SP/ NKA. Thus, SP, NKA, and CGRP are all present in sensory nerves in and under the airway epithelium (25), and their synthetic mechanisms have been established (28, 29). They are released from the nerve terminals by a Ca²⁺-dependent mechanism and diffuse through the mucosa. Local application of the agents mimics endogenous release. They are inactivated by local enzymes, especially NEP (30), and their actions are inhibited by NEP and are potentiated by NEP inhibitors. Specific antagonists are effective in blocking endogenous and exogenous effects of the peptides. Extensive research on the subject has been reviewed (31).

With regard to airway smooth muscle, most experiments have been with rodents, especially the guinea pig. Both SP and NKA are potent contractile agents of the smooth muscle, the latter being stronger, which points to an NK2 receptor mechanism (32). CGRP seems to have no action on bronchial smooth muscle, or even to cause a weak relaxation (33). Field stimulation causes a smooth muscle contraction that is only partly inhibited by atropine; in the case of vagal stimulation, atropine and the ganglionic blocking agent hexamethonium again only partly block the contractile response (34). Both the response and the release of the sensory neuropeptides can be inhibited by large doses of the neurotoxic agent capsaicin, which destroys sensory nerves or depletes them of their neuropeptides; in small concentrations capsaicin stimulates the sensory receptors, and it has been much used in the study of neurogenic inflammation.

In species other than rodents, evidence for an e-NANC control of airway smooth muscle is scanty or absent. Thus, in dogs and cats capsaicin causes a bronchoconstriction that is entirely blocked by atropine and is therefore presumably a vagal cholinergic reflex (17, 18). Indeed, as mentioned earlier, stimulation of lung C-fiber receptors can cause a bronchodilatation via the i-NANC system (19).

In humans there seems to be no evidence for an e-NANC contraction of airway smooth muscle (31). Field stimulation of bronchial preparations from healthy or diseased airways causes a bronchoconstriction that is completely blocked by atropine (35). Although capsaicin applied directly to human airways can cause a small bronchoconstriction (36), the concentrations of the agent have to be high and it is known not to be specific for C-fiber receptors. Inhaled capsaicin in humans causes bronchoconstriction (37), but this is blocked by atropine and is presumably a reflex. In spite of these negative results, SP or NKA applied to human airway smooth muscle causes contraction via NK2 receptors (32, 36), and inhalation of high concentrations of SP can also cause bronchoconstriction (38). Presumably the abundant presence of NEP in the airway epithelium would lessen the effect of inhaled tachykinins. There is some evidence that in subjects with asthma, when the epithelium may be damaged, there is bronchial hyperresponsiveness to inhaled tachykinins (38).

Further evidence that neurogenic inflammation is weak or absent in humans comes from studies on the nasal mucosa. In the guinea pig that was an early and classic tissue used to study neurogenic inflammation, which can be induced by application of capsaicin and is characterized by an exudate of liquid, including albumin. Application of capsaicin to the human nasal mucosa causes intense pain and sneezing, evidence that sensory nerves are being stimulated, but no exudate of albumin (39). The latter can be powerfully induced by application of histamine, which also causes sneezing but little pain (39).

A further action of the tachykinins is to potentiate cholinergic bronchoconstriction by an action at the ganglionic level (40), and possibly also at postganglionic vagal nerves (41). At the ganglionic level the association between SP and vagal motor cell bodies has been established histologically (7) and studied electrophysiologically (42). These studies have been with guinea pigs, and similar work with human tissue has not indicated any such mechanism.

	PROBLEMS

TOP ABSTRACT INTRODUCTION NONADRENERGIC/NONCHOLINERGIC NONADRENERGIC/NONCHOLINERGIC PROBLEMS REFERENCES

A summary of the vast body of research on the NANC systems to airway smooth muscle is that VIP and NO are established as i-NANC neurotransmitters, the latter being the only or main agent in humans, and that the tachykinins SP and NKA are the e-NANC neurotransmitters that mediate neurogenic inflammation in rodents, a process the existence of which is debatable in humans. However, it should be emphasized that it is impossible, for practical and ethical reasons, to reproduce in humans many of the experimental techniques used in other species.

The significance of cotransmission is perplexing. As already described, most VIP-containing nerves are also cholinergic, containing choline acetyltransferase, and many also contain PHI and/or PHM. VIP has also been shown to co-localize with NPY in nerves in sympathetic ganglia and in parasympathetic motor nerves, and with either enkephalin, SP, or CGRP in mucosal nerves, including probably in the epithelium. NO (assessed as iNOS) is present in cholinergic motor nerves together with VIP. On the other hand, many nerves do not seem to show co-localization, or perhaps the correct neurotransmitters have not been sought. A recent study of the ferret has shown that some motoneurones contain only one of ACh, VIP, iNOS, or SP, and others contain various pairs, triplets, or quadruplets of the neurotransmitters (9). SP, as well as being in some cell bodies, was also present in nerve fibers around various types of cells. Thus, different motor nerves contain different representations of the various neurotransmitters, but we do not know whether there are separate central nervous and reflex controls for the various types of nerves. In this respect, the observation that the i-NANC system to the airway smooth muscle of the cat can be activated reflexively by the pulmonary C-fiber reflex, but not by the reflex from peripheral chemoreceptors (19), indicates that there may be different controls for different motor pathways.

Presumably a nerve impulse will release quantities of all the co-localized neurotransmitters, and the balance of amounts released is thought to depend on the frequency of nerve impulses. In general the "conventional" neurotransmitters, ACh and norepinephrine (NE), have greater quantities released at lower firing frequencies, and the neuropeptides at higher frequencies. However, this frequency-dependence, although shown to exist in airway motor nerves, has scarcely been put on a quantitative basis. And the nerve excitations used to study frequency-dependence in field and nerve stimulation experiments are usually regular chains of impulses, and in no way resemble the natural firing pattern of either motor or sensory nerves (43). For the sensory nerves, both the amounts of the released neurotransmitters (SP, NKA, CGRP, and possibly VIP and NO) and whether their individual releases are frequency-dependent, have not been studied. For all the nerves, motor and sensory, there has been virtually no quantitation of the amounts of transmitter present per nerve, the amounts released per nerve impulse, and the resulting local tissue concentrations.

Equally puzzling must be the response of the target tissue to the mixture of neurotransmitters. Some may act in the same direction; examples are SP and NKA on airway smooth muscle and both with CGRP on blood flow and with NPY and NE on blood vessels. But others act in opposite directions; VIP and NO relax airway smooth muscle while ACh contracts it, and VIP and NPY in sympathetic nerves would have opposite actions on blood vessels. One possibility is that the different co-localized neurotransmitters act on different target organs. This is true for VIP and ACh in salivary and other glands (44), where the peptide is a vasodilator and ACh is a secretagogue. However, for the airways we lack the quantitative information to see if this possibility is likely. If two or more neurotransmitters with opposing actions are targeting the same tissue, for example, NO and ACh on airway smooth muscle, the response of the target seems unnecessarily complicated. It has been suggested that ACh contracts the muscle and that NO or VIP, with longer-lasting actions, would subsequently switch off the contraction. Thus, vagal stimulation causes a contraction that reverses to a relaxation (45). Such a mechanism has never been demonstrated with physiologic stimuli. In any event it raises the teleologic question as to what, if any, are the physiologic benefits of the existence of airway smooth muscle and of its contraction and relaxation, and what, if any, is the benefit of the contraction being increased in conditions such as asthma? The statement that "co-transmission represents a fundamental mechanism employed by autonomic nerves to achieve efficient and precise control over a target tissue" (10) seems, in the lack of evidence, rather wishful thinking. The motor system with the greatest efficiency and precision is that to skeletal muscle, for example, to control movements of the fingers, the vocal cords, and the eyes. Yet that control is achieved with only one neurotransmitter, ACh. At a more mundane level, one does not have a more efficient and precise control of a motor car if one is operating the accelerator and the brake at the same time, or of an army if one is ordering it to advance and retreat simultaneously. The analogies may not be perfect, but the benefits of opposing controls to airway smooth muscle need to be established. Although the functional existence of cotransmission seems indisputable, much more research is needed to demonstrate how it works and what evolutionary advantages it has provided in health and disease.

Most experimental studies have concentrated on one neurotransmitter and one target tissue, for obvious reasons. However, in vivo the situation is far more complicated. Not only is there co-localization and presumably co-release of transmitters, but secondary responses may cause changes in the release of other agents. As mentioned above, these may include NO and endothelins from blood vessel endothelium and various mediators from airway epithelium. The size and possibly the nature of the end responses will depend on the balance of action of all these mediators, and such complex interactions have scarcely been studied. A further complication applies to the sensory neuropeptides. As well as their local effects, the same nerves may set up airway reflexes, including bronchoconstriction, vasodilatation, and mucus secretion (17, 18). Furthermore, it has been shown that SP can stimulate RARs in the mucosa (46) as a secondary effect of plasma extravasation, and these receptors will in turn activate similar reflexes and also possibly cough. It is interesting that this mechanism seems to be NO-dependent (47).

A final consideration is the role of the NANC system in airway disease. There has been much discussion as to whether, for example, the bronchoconstriction of asthma may be due, at least in part, to downregulation of the VIP response of airway smooth muscle, and whether the same process may relate to NO in humans (4). The important concept of sensory nerve hyperresponsiveness has been evoked to explain bronchial muscle hyperresponsiveness and hypersensitive cough in asthma, and also similar mechanisms in rhinitis (48, 49). In view of the complexity of the NANC system mechanisms in healthy animals, it is perhaps not surprising that the roles of the system in disease, while very plausible, remain to be elucidated in detail (50, 51).