I. Introduction

Top Next References

Asthma is a complex chronic inflammatory disease of the airways that involves the activation of many inflammatory and structural cells, all of which release inflammatory mediators that result in the typical pathophysiological changes of asthma (Barnes, 1996a) (table 1). By inflammatory mediators, we mean cell products that are secreted and exert functional effects. We reviewed the mediators of asthma in 1988 (Barnes et al., 1988), but since then there have been major advances in our understanding of the mechanisms of asthma and the role of inflammatory mediators. There is now greater understanding of each mediator; in addition, novel mediators of asthma, such as the cytokines, have been identified. To date, >50 different mediators have been identified in asthma. Advances in this field have been greatly assisted by the development of potent and specific inhibitors that either block the inflammatory mediator receptors or inhibit mediator synthesis.

In writing this review, we have focused on new developments since 1988 and have emphasized studies in humans wherever possible. There is a vast and rapidly increasing body of literature on mediators of asthma; therefore, we have been forced to be somewhat selective. We have chosen to emphasize the mediators and effects that we think are most relevant to human asthma.

Many inflammatory cells are recruited to asthmatic airways or are activated in situ. These include mast cells, macrophages, eosinophils, T lymphocytes, dendritic cells, basophils, neutrophils, and platelets. It is now increasingly recognized that structural cells may also be important sources of mediators in asthma. Airway epithelial cells, smooth muscle cells, endothelial cells, and fibroblasts are all capable of synthesizing and releasing inflammatory mediators (Levine, 1995; Saunders et al., 1997; John et al., 1997). Indeed, these cells may become the major sources of inflammatory mediators in the airway, and this may explain how asthmatic inflammation persists even in the absence of activating stimuli.

There have been major advances in our understanding of the synthetic pathways involved in the synthesis of inflammatory mediators. Many of the key enzymes have now been cloned; in several cases, specific inhibitors have been developed that may have useful therapeutic effects. 5-Lipoxygenase (5-LO)^b inhibitors, which inhibit the synthesis of leukotrienes (LTs), have already been shown to have beneficial effects in the control of clinical asthma and are now available for clinical use (Israel et al., 1996).

Many inflammatory mediator receptors have now been cloned. The receptor for platelet-activating factor (PAF) was the first inflammatory mediator receptor to be cloned (Honda et al., 1991), and many inflammatory mediator receptors have been sequenced since then. The receptors for many inflammatory mediators have the typical seven-transmembrane domain structure that is expected for G protein-coupled receptors. However, receptors for cytokines and growth factors have markedly different structures, and usually two or more subunits are involved (Kishimoto et al., 1994). Receptor cloning has yielded a much better understanding of receptor function, because the receptors can be expressed in cell lines, allowing investigation of the "pure" pharmacological features of the receptor and enabling screening for drugs that interact with the receptor. This has been important in elucidating the signal transduction pathways involved in receptor function. Many signal transduction pathways have now been identified. For noncytokine mediators, inflammatory receptors are often coupled, through G proteins (G_q and G_i), to phosphoinositide (PI) hydrolysis, but it is increasingly recognized that other pathways may also be activated, including the complex mitogen-activated protein (MAP) kinase pathways that are involved in more long term effects of mediators. Cytokine receptors signal through complex pathways, including MAP kinases and other protein kinases, that result in the activation of transcription factors. Transcription factors regulate the expression of many genes, including inflammatory genes themselves.

The cloning of receptors has made it possible to study the factors regulating their expression. This may be of particular relevance in asthma, because the inflammatory state may alter the gene expression, translation, or function of receptors, thus affecting responsiveness to different mediators.

Inflammatory mediators produce many effects in the airways, including bronchoconstriction, plasma exudation, mucus secretion, neural effects, and attraction and activation of inflammatory cells. Although the acute effects of mediators have been emphasized, there is increasing recognition that mediators may result in long-lasting structural changes in the airways that are also mediated by the release of inflammatory mediators. These changes may include fibrosis resulting from the deposition of collagen, which is seen predominantly under the epithelium even in patients with mild asthma. The airway smooth muscle layer is also thickened in asthma, and this is likely the result of increases in the number of smooth muscle cells (hyperplasia) and increases in their size (hypertrophy) (Knox, 1994). There may be proliferation of airway vessels (angiogenesis) (Kuwano et al., 1993) and of mucus-secreting cells. There may also be changes in the innervation of the airways.

There are several lines of evidence that may implicate a mediator in asthma. Firstly, it may mimic features of clinical asthma. Secondly, the mediator may be produced in asthmatic patients. Thus, mediators or their metabolites may be detected in plasma (e.g., histamine), urine (e.g., LTE₄), or, more likely, the airways (in biopsies, bronchoalveolar lavage fluid, induced sputum, or exhaled air). However, this does not necessarily mean that the mediator plays any important role in asthma. The best evidence for the involvement of a mediator in asthma is obtained with the use of specific blockers. These may be drugs that block the synthesis of the mediators (e.g., 5-LO inhibitors) or drugs that block their receptors (e.g., antihistamines). Use of new and selective mediator blockers has enormously increased our understanding of the individual mediators and also of asthma itself. Although it is unlikely that blockade of a single mediator will be entirely effective in controlling asthma, there is accumulating evidence that some mediators are more important than others. PAF receptor antagonists are of no obvious clinical benefit in asthma (Kuitert et al., 1993), but cysteinyl-LT (cys-LT) receptor antagonists have considerable clinical effects (O'Byrne et al., 1997).

The role of a mediator in asthma may be difficult to assess when the mediator has a long term effect on airway function. It is easy to measure the effect of a mediator on airway smooth muscle, but it is more difficult to determine its effect on airway microvascular leakage and mucus secretion. It may be even more difficult to determine the role of a mediator on chronic inflammatory effects, such as airway smooth muscle proliferation and fibrosis, that may develop over many years. However, prevention of the long term consequences of asthmatic inflammation, such as irreversible airway narrowing, may be an important goal of asthma therapy, and it is necessary to devise methods to investigate how mediators may affect these long term consequences of asthma.

Asthma has a characteristic clinical pattern, and the histological appearance of asthma is very similar among patients, even when there are differences in asthma severity or in whether or not the asthma is allergic. However, it is likely that there are differing mechanisms of asthma among patients and that different patterns of inflammatory mediators are involved. This suggests that mediator antagonists would have different effects in different patients. This has already been observed in the use of anti-LTs, because some patients appear to have much better therapeutic responses than others. This might be related to polymorphisms of the 5-LO gene (In et al., 1997), but there might be differences that relate to the type of asthma. Patients with aspirinsensitive asthma are particularly helped by anti-LTs, consistent with a critical role for cys-LTs in this type of asthma. As more mediator antagonists become available, other patients who respond well to a particular antagonist may be identified and the heterogeneity of asthma may be revealed.

Although in the past much attention has been paid to acute inflammatory responses (such as bronchoconstriction, plasma exudation, and mucus hypersecretion) in asthma, it is being increasingly recognized that chronic inflammation is an important aspect of asthma (Redington and Howarth, 1997). This chronic inflammation may result in structural changes in the airway, such as fibrosis (particularly under the epithelium), increased thickness of the airway smooth muscle layer (hyperplasia and hypertrophy), hyperplasia of mucus-secreting cells, and new vessel formation (angiogenesis). Some of these changes may be irreversible, leading to fixed narrowing of the airways. These chronic inflammatory changes are mediated by the secretion of distinct mediators, although their role in asthma is still far from certain. These factors include cytokines and growth factors. Cytokines are a large group of protein mediators that play a critical role in determining the nature of the inflammatory response and its persistence. They play a key role in the pathophysiological changes in chronic asthma and are being increasingly recognized as important targets for treatment (Robinson et al., 1993c; Barnes, 1994a; Drazen et al., 1996).

Transcription factors are DNA-binding proteins that regulate the expression of inflammatory genes, including enzymes involved in the synthesis of inflammatory mediators and protein and peptide mediators. Transcription factors therefore play a critical role in the expression of inflammatory proteins in asthma, because many of these proteins are regulated at a transcriptional level (Barnes and Karin, 1997; Barnes and Adcock, 1998). These transcription factors include nuclear factor-B (NF-B) and activator protein-1 (AP-1), which are universal transcription factors that are involved in the expression of multiple inflammatory and immune genes and may play a key role in amplifying the inflammatory response. Other transcription factors, such as nuclear factor of activated T cells (NF-AT), are more specific and regulate the expression of a restricted set of genes in particular types of cell; NF-AT regulates the expression of interleukin (IL)-2 and IL-5 in T lymphocytes.

	II. Amine Mediators

Top Previous Next References

1. Synthesis and metabolism. Histamine is synthesized and released by mast cells in the airway wall and by circulating and infiltrating basophils. Although airway mast cells are likely to be the major cellular source of histamine in asthma, there is increasing evidence that basophils may be recruited to asthmatic airways and may release histamine in response to cytokine histamine-releasing factors (Schroeder and MacGlashan, 1997).

2. Receptors. Histamine has multiple effects on airway function that are mediated by specific surface receptors on target cells (Barnes, 1991). Three types of histamine receptors have now been recognized pharmacologically (Hill, 1990). Histamine receptors were first differentiated into H₁ and H₂ receptors by Ash and Schild in 1966, when it was found that some responses to histamine were blocked by low doses of mepyramine (pyrilamine), whereas others were insensitive. This classification was supported by the development of H₂ receptor-selective antagonists, such as cimetidine and ranitidine. Both H₁ and H₂ receptors have been cloned. Both have the seven-transmembrane domain motif typical of G proteincoupled receptors. A third histamine receptor subtype, termed H₃, has been described more recently; this receptor acts as an inhibitory autoreceptor in the central nervous system (Schultz et al., 1991).

3. Effects on airways. a. AIRWAY SMOOTH MUSCLE. Histamine stimulates PI hydrolysis in airway smooth muscle (Grandordy and Barnes, 1987; Hall and Hill, 1988; Daykin et al., 1993), and there is a close association of receptor occupancy, PI hydrolysis, and the contractile response, indicating that there are few or no "spare" receptors (Grandordy and Barnes, 1987). Histamine also increases the concentration of inositol-1,4,5-trisphosphate (IP₃) in airway smooth muscle, although the magnitude of the increase is less than with cholinergic agonists, which may reflect lower receptor density (Chilvers et al., 1989). In cultured human airway smooth muscle cells, histamine increases [Ca²⁺]_i via an increase in IP₃ levels (Hardy et al., 1996).

d. NERVES. In many species, the bronchoconstricting effect of histamine is partially mediated by a vagal cholinergic reflex and may be modulated by muscarinic receptor antagonists. In dogs, histamine increases the discharge of "irritant" receptors in vivo (A-fibers), and these effects are abolished by H₁ antagonists. However, in vitro measurements of single afferent fibers in guinea pig trachea show no evidence for activation of either A- or C-fibers by histamine (Fox et al., 1993

4. Role in asthma. a. RELEASE. Measurement of histamine in the circulation is complicated by the spontaneous release from basophils, and measurement of stable metabolites in the urine may not reflect release from mast cells in the airways. Several previous studies demonstrated elevations of plasma histamine concentrations in patients with asthma, at rest, after exercise, at night, and after allergen challenge, but these studies are difficult to interpret because of the likelihood of contamination from basophil release in the collected blood samples (Ind et al., 1983). It is possible that basophils from patients with asthma may be more "leaky" and that this may contribute to the higher concentrations measured in asthmatic patients. Studies of histamine infusions in normal volunteers have demonstrated that doses of histamine that yield the plasma concentrations reported in patients with asthma have marked cardiovascular effects, indicating that the higher levels seen in the blood of asthmatic patients are likely to be generated in vitro during storage and preparation of the plasma samples. The histamine released from the airways may increase plasma concentrations, but this may be overwhelmed by the contribution from circulating basophils. Sampling closer to the site of histamine release may overcome these problems. Venous sampling in the arm shows an increase in plasma histamine concentrations after mast cell degranulation in the skin of the arm, induced by SP (Barnes et al., 1986), but such sampling is not feasible in the airways. Measurement of histamine in bronchoalveolar lavage fluid is likely to provide a much more direct measurement of airway histamine release. There is evidence that histamine concentrations are elevated in bronchoalveolar lavage fluid of asthmatic patients, both at rest and after allergen challenge (Liu et al., 1991; Wenzel et al., 1988). The source of histamine is presumed to be mucosal mast cells, and the contribution of infiltrating basophils is unclear.

Serotonin [5-hydroxytryptamine (5-HT)] causes bronchoconstriction in most animal species, but interest in this mediator is minimal because it is not a constrictor of human airways and its relevance in asthma seems doubtful (Barnes et al., 1988).

2. Receptors. Multiple serotonin receptors have now been recognized, based on the development of selective antagonists and molecular cloning (Saxena, 1995). There are up to seven types of 5-HT receptors, each with several subtypes. Selective antagonists have now become available for clinical use, but few have been used in investigations of human airway cells or in the treatment of patients with asthma.

3. Effects on airways. Serotonin does not constrict human airway smooth muscle in vitro and may even have bronchodilating effects, although pulmonary vessels are constricted as expected (Raffestin et al., 1985). In animals, serotonin increases acetylcholine release from airway nerves, and this has been demonstrated in human airways (Takahashi et al., 1995). The receptor mediating this response appears to be a 5-HT₃ receptor (Takahashi et al., 1995). In guinea pig airways, serotonin inhibits nonadrenergic noncholinergic (NANC), neurally induced constriction resulting from tachykinin release via a 5-HT₁-like receptor localized to sensory nerve endings (Ward et al., 1994; Dupont et al., 1996). In humans, infused serotonin has no effect on airway function but may have an inhibitory effect on cough reflexes, possibly mediated by receptors on airway sensory nerves (Stone et al., 1993). Serotonin is a potent inducer of microvascular leakage in rodent airways, but it is not certain whether serotonin has this property in human airways. Serotonin has a blocking effect on sodium channels in human airway epithelial cells, but the receptor subtype involved has not been established (Graham et al., 1992).

4. Role in asthma. Plasma serotonin levels are reported to be elevated in asthma and are significantly related to asthma severity (Lechin et al., 1996). The source of serotonin is likely to be platelets, but the clinical relevance of this observation is unclear.

C. Adenosine

1. Synthesis and metabolism. Adenosine is a purine nucleoside that is produced by dephosphorylation of 5'-AMP by the membrane-associated enzyme 5'-nucleotidase and is liberated intracellularly by cleavage of the high energy bonds of adenosine triphosphate, adenosine diphosphate, and cyclic 5'-AMP. However, during hypoxia or even excessive cell stimulation, when the utilization of energy and oxygen exceeds the supply, 5'-AMP is metabolized to adenosine (Mentzer et al., 1975). This conversion is performed by extracellular 5'-nucleotidase. Adenosine release was originally demonstrated during myocardial hypoxia (Mentzer et al., 1975), although there is now evidence that all cells are capable of producing adenosine in times of energy deficit. Adenosine can be released by lung tissue in times of hypoxia, such as after allergen-induced bronchoconstriction, when the circulating levels of adenosine have been shown to be 3 times the base-line concentrations (Mann et al., 1986). Mast cells are a likely source of adenosine in this situation, because these cells have been shown to be capable of releasing adenosine in response to IgE cross-linking and other stimuli for mast cell activation (Marquardt et al., 1986).

2. Receptors. Three distinct subtypes of receptor have been characterized to date, based on biochemical, functional, and more recent cloning studies (Linden et al., 1991; Linden, 1994). These receptors include the A₁, A_2a, A_2b, and A₃ receptor subtypes. Interaction of adenosine with these receptors leads to either inhibition of adenylyl cyclase (A₁), stimulation of adenylyl cyclase (A_2a and A_2b) (Collis and Hourani, 1993), or activation of phospholipase C (A₃) (Ali et al., 1990). The A₁ receptor is expressed in lung tissue (Ren and Stiles, 1994) and, in particular, A₁ receptors have been identified on human epithelial cells (McCoy et al., 1995). The classification of adenosine receptors into A_2a and A_2b subtypes is based on distinct rank orders of potency of a range of agonists and antagonists and distinct nucleotide sequences of the two complementary deoxyribonucleic acids (cDNAs). A_2a, A_2b, and A₃ receptors are expressed in several tissues, including lungs, and in mast cells and fibroblasts (Linden et al., 1993; Auchampach et al., 1997; Ciruela et al., 1997; Shryock and Belardinelli, 1997; Fredholm, 1997).

3. Effects on airways. a. AIRWAY SMOOTH MUSCLE. Adenosine elicits little or no contraction of human bronchi from nonasthmatic subjects but potently constricts asthmatic airways in vitro (Björck et al., 1992). This constriction is blocked by histamine and LT antagonists and is therefore likely to be attributable to the release of mediators from mast cells in asthmatic airways. It is likely that the bronchoconstricting effects of adenosine are indirect, resulting from the activation of mast cell degranulation, because adenosine causes histamine release from mast cells (Church et al., 1986). Comparable results have been observed in vivo, where adenosine and AMP are able to elicit bronchoconstricting effects in atopic and asthmatic subjects but have no effect in normal subjects (Cushley et al., 1983). Furthermore, dipyridamole (an inhibitor of adenosine uptake into tissues) enhances adenosine-induced bronchospasm in asthmatic subjects (Crimi et al., 1988), an effect that can be inhibited by theophylline (a nonselective adenosine antagonist) (Cushley et al., 1984). The receptor mediating the bronchoconstricting effect of adenosine in asthma is not yet known. In rabbits, the A₁ receptor is a likely candidate, because tracheal strips from rabbits immunized with house dust mites are more responsive to adenosine and the adenosine A₁-selective agonist cyclopentyladenosine than are tracheal strips isolated from naive animals (Ali et al., 1994a). Furthermore, immunized animals are considerably more responsive to the bronchoconstricting effects of adenosine (Thorne and Broadley, 1994) and cyclopentyladenosine in vivo (Ali et al., 1994b; el Hashim et al., 1996). No bronchoconstricting effects of the A₃-selective agonist aminophenylethyladenosine have been found in rabbits (el Hashim et al., 1996) or guinea pigs (Hannon et al., 1995), although studies in rats have shown that aminophenylethyladenosine can elicit bronchoconstriction (Meade et al., 1996). The histamine-releasing effect of adenosine may involve the A_2b receptor, because this effect is sensitive to enprofylline (an A_2b receptor antagonist) (Feoktistov and Biaggioni, 1995). Certainly, there is clinical evidence showing that elevated levels of histamine can be demonstrated in plasma after the inhalation of AMP by atopic subjects (Phillips et al., 1990), and increased levels of histamine have been detected after the instillation of AMP directly into the airways (Polosa et al., 1995). Furthermore, the H₁ receptor antagonist terfenadine has a protective effect against adenosine-induced bronchoconstriction in asthmatic subjects (Rafferty et al., 1987).

4. Role in asthma. a. RELEASE. Increased levels of adenosine have been found in bronchoalveolar lavage fluid obtained from asthmatic subjects, compared with normal subjects (Driver et al., 1993), and, as discussed above, adenosine concentrations in plasma are higher in allergic patients minutes after allergen provocation (Mann et al., 1986). A₃ receptor expression is increased in asthmatic lungs, compared with lungs of normal subjects, although, because the A₃ receptor is expressed predominantly in eosinophils, this may be a reflection of eosinophilic infiltration (Walker et al., 1997).

	III. Lipid-Derived Mediators

Top Previous Next References

1. Synthesis and metabolism. Prostanoids are generated from arachidonic acid by two forms of COX (Mitchell et al., 1995). COX-1 is constitutive and is responsible for basal release of prostanoids, whereas COX-2 is inducible by inflammatory stimuli, such as endotoxin and proinflammatory cytokines, and its induction is inhibited by glucocorticoids. Both COX-1 and COX-2 are expressed in human lung (Demoly et al., 1997). Human airway epithelial cells basally express COX-1, whereas COX-2 is induced by IL-1 and TNF- (Mitchell et al., 1994; Newton et al., 1997b; Asano et al., 1997) and is enhanced by NO (Watkins et al., 1997). COX-2 is also induced in cultured human airway smooth muscle cells by proinflammatory cytokines and bradykinin (Belvisi et al., 1997; Pang and Knox, 1997a,b), and the formation of prostanoids is blocked by the selective COX-2 inhibitor L745,337 (Saunders et al., 1998). COX-2 expression is inhibited by dexamethasone in both epithelial cells and smooth muscle cells. The induction of COX-2 is regulated in part by NF-B, and this may also account for the inhibitory action of glucocorticoids (Newton et al., 1997a). There is no difference in the profiles of prostanoids formed by COX-1 and COX-2. In epithelial cells and airway smooth muscle cells, the predominant prostanoids are PGE₂ and 6-keto-PGF₁ (metabolite of PGI₂), whereas there is relatively little formation of Tx (Mitchell et al., 1994; Belvisi et al., 1997). Tx is formed from the intermediate PGH₂ by a distinct enzyme, Tx synthase, which has been cloned (Ohashi et al., 1992).

2

2. Receptors. Several prostanoid receptors have now been cloned (Ushikubi et al., 1995; Pierce et al., 1995). Pharmacologically, prostanoid receptors are classified according to the prostanoid that causes selective activation; PGE₂ preferentially activates EP receptors, PGI₂ (prostacyclin) activates IP receptors, PGF₂ activates FP receptors, PGD₂ activates DP receptors, and Tx activates TP receptors (Coleman et al., 1994). Within each receptor type there may be distinct subtypes, many of which have been identified using selective ligands and cloning; the EP receptor has at least four subtypes, which are differentially expressed in different cell types. EP₁ receptors mediate activation responses and are involved in hyperalgesic responses, whereas EP₂ and EP₄ receptors mediate smooth muscle relaxation responses and EP₃ receptors modulate neurotransmitter release. In airway smooth muscle, several constrictor PGs (PGD₂, PGF₂, and 8-epi-PGF₂) appear to work through activation of TP receptors (Coleman and Sheldrick, 1989; Kawikova et al., 1996).

3. Effects on airways. a. AIRWAY SMOOTH MUSCLE.