PATHOGENETIC MECHANISMS

Contents

This document was last modified on 21-May-2009.

Introduction

A bewildering number of hypotheses about the pathogenesis of asthma have been suggested this century, most of which are supported by at least some evidence. The multiplicity of aetiological agents reviewed earlier imply completely different mechanisms of causality at the biological level. Broadly, these may be divided into allergic, inflammatory, neurogenic, and physical mechanisms. Evidence is most extensive currently for a combination of allergic and inflammatory mechanisms. I will organise this review by first reviewing the histopathology of asthma, concentrating on the various cell types associated with active disease, and then the immunology and biochemistry of allergy and of asthma.

Pathology of the asthmatic lung

The airway obstruction in acute fatal asthma arises through bronchial smooth muscle contraction, airway wall thickening and intraluminal mucus and debris. The wall thickening is characterised by submucosal oedema, vasodilatation, with cellular infiltrates, predominantly lymphocytes and eosinophils. Smooth muscle hypertrophy and oedema is present. The epithelium is also infiltrated by cells, and is sloughy, leaving patches of denuded basement membrane. The numbers of mucosal goblet cells, and submucosal mucous glands is increased. The lumen contains mucus plugs, shed epithelium (Creola bodies) and Charcot-Leyden crystals (crystallised eosinophil derived major basic protein). Mucus is relatively dehydrated, and so viscid.

Bronchoscopic biopsies taken from mild and asymptomatic asthmatics find evidence of similar inflammatory processes. Significantly more active eosinophils are present in the submucosae of mild and even asymptomatic asthmatics compared to controls. T lymphocytes are also present in increased numbers. Numbers of mast cells and neutrophils are not usually increased [Djukanovic et al 1990]. There is thickening of the basement membrane, which electron microscopy has revealed to be due to deposition of (banded) collagen by myofibroblasts. Lavage fluid from the lumen of the asthmatic airway contains numerous mediators including histamine, prostaglandins (predominantly PGD2 and PGF2- alpha), leukotrienes, kinins, kallikrein, and eosinophil derived major basic protein.

Cells involved in the pathogenesis of asthma

In the current model of asthma as an allergic disease, two cell types are central. These are the lymphocyte, especially in the role of T- cell as a regulator of immune processes, and the eosinophil, as effector cell causing tissue damage. The mast cell retains an important role, but I will defer discussion. Increasing importance is also being attached to specialised antigen presenting cells (APCs) such as the dendritic cell.

T lymphocytes

In current models of immune regulation, T-cells determine the presence or absence as well as the nature of the immune response the body mounts to various stimuli by release of various patterns of cytokines. A number of lines of evidence now suggest that T-cells play a central role in the development and perpetuation of asthma and atopy.

Okumura et al [1971] were the first workers to show that IgE production is dependent on and is regulated by the presence of T-cells. These are mainly helper T (TH) cells (CD3+, CD4+), which express the low affinity IgE receptor (FcEpsRII or CD23), and produce a variety of substances such as IgE binding factors (which include soluble fragments of the FcEpsRII molecule), and lymphokines that can suppress or increase IgE production by B lymphocytes and plasma cells.

Burastero and coworkers [1993] showed that the peripheral blood level of D pter-specific T-cells was elevated in asthmatics compared to controls, and that this level was correlated with degree of bronchial responsiveness to inhalant challenge with this allergen. Walker et al [1991] found increased numbers of IL-2 receptor (CD25) bearing T-cells in the peripheral blood of atopic and nonatopic asthmatics compared to nonatopic controls. It should be noted that the nonatopic asthmatics were older, and although skin-test negative, a sIgE over 95 IU/ml was seen in five out of the ten.

A number of studies utilising bronchial alveolar lavage have demonstrated the presence of active T-cells in the airways of asthmatics. Robinson et al [1992] found no increase in numbers of lymphocytes recovered, but did find significantly increased numbers of cells expressing mRNA for GM-CSF, IL-2, IL-3, IL-4 and IL-5, which were revealed to be T-cells on double fluorescent labelling for the T-cell receptor (CD3). Several studies [Kay 1992] have demonstrated abrupt falls in numbers of peripheral blood CD4+ T-cells following allergen inhalation challenge, which mirror increases in numbers of T-cells in lung lavage fluid. Cho et al [2005] showed T-cells from asthmatic sputum produce more IL-4, IL-5 and IFN-gamma than those from control sputum along a disease severity gradient.

The T-cells involved in atopy exhibit immunological memory. Memory T-cells differ from virgin T-cells in that they express specific cell surface molecules: an alternative form of CD45 -- CD45R- (to which a monoclonal antibody exists); increased levels of CD2, LFA-1 and PgP- 1 -- involved in cell adhesion; and the IL-2 receptor [Vitteta et al 1991]. They produce increased levels of particular lymphokines -- IL2, and IL4 or IFN-gamma. If the appropriate allergen is presented, the T-cell will undergo clonal proliferation. Although HDM specific clones can be raised from nonatopic donors [O'Hehir et al 1991], those from sensitised atopic donors are self-supporting.

An elegant quasiexperiment demonstrating the role of these cells was reported by Agosti et al [1988] and Higgenbottam and Varma [1989], where skin atopy and asthma were transmitted by allogeneic bone marrow transplantation, the effects lasting over one year. In the former study, 12 nonatopic patients receiving marrow from atopic donors were selected. In 8 out of 11 of these (that had survived to one year), the recipient developed positive skin prick tests to allergens to which the donor was sensitised. By contrast, they observed no responses to allergens where the donor was unreactive. RAST for mite-specific IgE showed long term survival of a donor-derived allergic response in transplant recipients.

Eosinophils

Eosinophils are granulocytes whose cytoplasm contains eosin staining granules [Gleich 1991]. Production in the marrow is modulated by a number of lymphokines including IL-3,5,6, and GM-CSF. As mentioned earlier, eosinophils are currently regarded as the effector cells responsible for much of the pathology of asthma. It has been long known that atopy in general and asthma in particular are associated with blood and lung eosinophilia. The main role of the eosinophil seems to be a cytotoxic one, and it acts effectively both against host cells and parasites such as protozoa. Indeed, the "normal" state associated with both high levels of circulating IgE and eosinophils is parasitosis, where this combination of humoral and cellular host defences is thought most effective.

A point about the relationship between eosinophils and atopy in the literature should be noted. Eosinophils are important in diseases where no role for IgE has been delineated - such as "nonallergic" asthmas including that due to sensitisation to isocyanate [Paggiaro et al, 1990], aspirin, and other small molecules, as well as intrinsic asthma. As a result, however, some authors have argued (Chapter 1) that therefore occult allergic - IgE mediated mechanisms - must be present (perhaps early in their development) in these disease states.

Epidemiology and pathophysiology of eosinophilia

A number of studies have shown that peripheral blood eosinophilia is associated with asthma (and atopy) at all ages. For example, Tollerud et al [1989] describe the association in the Normative Aging Study where they found that the presence of eosinophilia (defined as >275 cells/mm3) was increased in asthma (OR=1.9), wheeze (OR=1.8), positive skin allergen testing (OR=3.4), and also chronic bronchitis and chronic productive cough (OR=1.9). They infer from the latter two associations that eosinophilia is a measure of airway inflammation, rather than just IgE mediated disease. Other workers such as Horn et al [1975], Bousquet [1990] and Griffin et al [1991] have shown that the degree of eosinophilia is proportional to the severity of asthma, as measured by clinical grading or pulmonary function. Specific bronchial challenge (both experimental and nonexperimental) in sensitive individuals is associated with a short term rise in eosinophil count. Large numbers of bronchial submucosal and lumenal eosinophils are pathognomic in status asthmaticus (acute severe asthma refractory to therapy). Eosinophils are increased in bronchoalveolar lavage fluid from asthmatics compared to nonasthmatics [DeMonchy et al 1985; Gleich 1990], as are subepithelial (and to a lesser extent deeper submucosal) eosinophils on bronchial biopsy [Djukanovic et al 1990; Bousquet et al 1990; Beasley et al 1989]. The technique of sputum induction by inhalation of hypertonic saline is a less invasive method of assessing airway eosinophil numbers, and has demonstrated similar correlations with asthma symptoms, BHR, and even with male sex in children [Gibson et al 1998].

Control of eosinophils

Since eosinophils do not express the FcEpsRI receptor, they must be recruited into both allergic and nonallergic inflammation by other mechanisms. This is probably mainly by means of lymphokines, especially IL-5 [Colley 1973], produced both by T-cells and by mast cells. Eosinophils do express the FcEpsRII receptor [Frew & Kay 1990], and this may be a less important source of recruitment into IgE mediated processes. IL-5 is known to mediate the transformation of eosinophils to the hypodense or activated form, which is seen in inflammation. In a recent study of fifteen asthmatics experiencing a moderate to severe exacerbation [Corrigan et al 1993], serum IL-5 levels were elevated sufficiently to be detectable in eight, but in none out of the seven controls. After one week of treatment with oral prednisolone, IL-5 was undetectable in any of the asthmatics. Walker et al [1991] found a correlation of 0.8 between eosinophilia (as well as eosinophil viability in vivo) and the number of IL-2 receptor bearing T- cells in the peripheral blood. As noted earlier, IL-2 receptor expression is a marker of an active/memory T-cell. Another cytokine that may be involved is IL-4, which leads to eosinophilia in mouse models. IgA is known to enhance eosinophil degranulation [O'Hehir et al 1991].

Eosinophil products

These cells secrete a number of proteins. The most important of these is major basic protein (MBP). This 31 kD protein is highly basic, and negatively charged in solution, and exhibits potent cytotoxic activity [Gleich 1990]. In vitro, MBP causes shedding of respiratory epithelium. In vivo, levels of MBP in the sputum of asthmatics reflect severity of current disease, while immunofluorescent studies show that MBP is present in the respiratory epithelial defects seen in asthmatics' bronchi.

Eosinophilic cationic protein (ECP) and eosinophil-derived neurotoxin (18-21 kD) are highly homologous proteins and exhibit ribonuclease activity (the latter more so). Both lead to exfoliation of respiratory epithelium. Both are highly basic. ECP also stimulates mast cell degranulation, inhibits T-cell activity, and shortens the coagulation time. ECP has been also found at increased levels in (late phase) BAL fluid from allergen challenged asthmatics. The secretory form of ECP within the eosinophil is recognised by the monoclonal antibody EG2, used therefore to identify activated eosinophils [Tai et al 1984]. Eosinophil peroxidase is cytotoxic in itself, but more so in the presence of hydrogen peroxide and halides. It too can cause mast cell degranulation.

Allergy and IgE

Description of IgE

Immunoglobulin E (for Erythema) is a glycoprotein made up of a distinguishing constant () region of four domains, and a variable region. The third constant domain C3 contains the separate binding sites for the high (1) and low (2) affinity IgE receptors [Sutton & Gould 1993]. McDonald et al [1987] have suggested that IgE produced by atopic subjects differs from that produced by nonatopics in being a more potent stimulus for basophil degranulation.

Cells associated with IgE production

T lymphocyte regulation of IgE production

This area has been the subject of several recent reviews [Lichtenstein & MacGlashan 1990; Romagnani 1990]. Interleukin 4 (IL-4) is a most important lymphokine regulating IgE. It is mainly produced by TH cells, as well as CD4- T cells and other inflammatory cells such as the mast cells [Plaut et al 1989]. The primary as well as much of the secondary IgE response is IL-4 dependent. IL-4, in the presence of direct T-cell/B- cell contact (and factors such as low molecular weight B-cell growth factor, and to a lesser extent other lymphokines) causes IgM producing B cells to class switch to IgE production. These B-cells may be a particular subclass able to produce IgM, IgD, or IgE, as the situation demands. It is important to note that a second signal such as T cell-B cell contact is required for IL-4 to be effective - either cognate (T cell receptor/allergen-CD3) or noncognate (possibly CD40) [Ricci 1992].

Gamma and alpha interferon (IFN-gamma, IFN-alpha) are TH-cell produced, potent inhibitors of IgE production. They are actively involved in the regulation of the IgE response, and cross-inhibit IL-4 production, and in mice, increase the IFN-gamma TH1-cell population while diminishing the IL-4 producing TH2 subgroup. They block IL-4 induced IgE synthesis at the mRNA transcription level in human B-cells [Gauchat et al 1991]. TH2-cells also produce IL-10, which inhibits TH1-cells indirectly, by reducing their response to APCs [Fitch et al 1993]. IL-12 is another lymphokine (produced by macrophages), that enhances TH1 proliferation [Hsieh et al 1993].

In culture systems, human TH-cells have been found to be predominantly IL-4 secreting if they recognised important allergens (TH2- like, Th2), and IL-2 and IFN-gamma secreting if specific to nonallergens such as tetanus toxoid [Wieranga et al 1990]. Kapsenberg et al [1991] provide even more specific findings. In this study, HDM specific T-cell clones from atopic individuals were shown to produce IL-4 solely, but those from non-atopics produced IFN-gamma, and only small amounts of IL-4. Non- allergen specific T-cell clones from the atopic subjects were IFN-gamma producers. Abnormalities of IL-4 and IFN-gamma have been reported in atopic dermatitis, with decreased levels of IFN-gamma production by mononuclear cells, and decreased proliferative response to IL-4. In clinical studies of atopic disease, IFN-gamma has shown great promise [Boguniewicz 1988]. TGF-beta similarly blocks IL-4 induced eta heavy chain transcription. IL-6 has also been shown to facilitate IgE synthesis.

The same mechanisms are functioning in the elevation of sIgE in helminth infection in (nonatopic) humans. In schistomiasis-infected subjects, King et al [1993] found parasite antigens (but not tetanus toxoid) stimulated lymphocytes IL-4 production in those patients exhibiting high serum parasite-specific IgE levels, and that those individuals with low or medium elevations had increased ratios of IFN to IL-4 producing CD4+ cells.

B lymphocytes

These will be from two broad groups, IgE memory cells, which will be independent of IL-4 regulation and the previously mentioned IgM secreting B-cells. Memory B-cells are generated in the germinal centres of lymph nodes and spleen, but long term memory cells tend to settle in the bone marrow [Vitteta et al 1991]. They differ from naive B-cells in expressing different surface immunoglobulins, such as sIgG and sIgA (as opposed to sIgM and sIgD) and have increased affinity (affinity maturation) for their specific antigen following somatic hypermutation and alteration in VH gene usage. Current theories see overproducing IgE B-cells as the victims of T-cell dysfunction, in view of their tight regulation by T-cells. IgE producing B- cells and plasma cells occur in largest numbers at mucosal surfaces in lung, gut and skin [Sutton & Gould 1993].

Effects of release of IgE

High affinity IgE receptor

Mast cells and basophils are the only effector cells that carry the high affinity IgE receptor (FcEpsRI), though it may be present in smaller numbers on antigen presenting cells (APCs) such as dendritic or Langherhans cells [Bieber et al 1992]. It is the crosslinking and aggregation of the receptor on the surfaces of these cells that leads to their activation and release of the major mediators of the allergic reaction.

FcEpsRI is a tetrameric glycoprotein. It comprises an alpha chain, beta chain and two disulphide linked gamma chains, all associated via non-covalent bonds. The alpha and gamma chain genes are both located on chromosome 1 (q21-23), and closely linked, possibly due to selection pressures. The beta gene (FCER1B) is located on chromosome 11q13. Since binding of IgE to the receptor is passive, it is the aggregation of receptor molecules that is essential to the activation of a calcium channel. One group [Hemmerich et al 1988] have tentatively identified the channel protein as the 89 kD cromolyn-binding protein - the target site for both cromoglycate and nedocromil.

Mast cells

Numbers of mast cells (specifically the mucosal type) are also controlled by IL-3 and IL-4. Schleimer [1990] notes that mucosal type mast cells are absent from mucosae in T cell deficient patients, and that in nude mice, a similar defect responds to IL- 3. IL-3 also is known to be a potent mast cell growth factor [O'Hehir et al 1991]. However, in vitro IL-3 and IL-4 stimulation of human mast cells in the absence of tissue fibroblast co-culture leads to transformation into basophils [Church et al 1991], and it has been suggested human mast cells lack the IL-3 receptor.

Activation of the mast cell high affinity IgE receptor causes calcium mediated exocytosis of the cytoplasmic granules, releasing histamine and proteoglycans such as heparin and chondroitin sulphate over a six minute period [Church et al 1991]. At the same time, synthesis and release of prostaglandins (mainly PGD2) and leukotrienes (mainly LTC4) is stimulated [Holtzman 1991].

An interesting recent finding is that FcEpsRI activation leads to mast cell production of lymphokines: IL-1,3,4,5 & 6, IFNg & GM CSF (a TH2 like pattern tends to predominate). IL-5 is an activator of eosinophils, so this might represent one mechanism for the recruitment of eosinophils to the sites of allergic reactions. Galli [1993] presents evidence from a mouse model that mast cell derived TNF-alpha is another mediator that can drive much of the cellular infiltration seen in early and late phase allergic reactions. These mediators are not released immediately, but must be synthesised - this would delay release to the time of the classic late allergic reaction. Since mast cells produce IL-3 and 4 on stimulation, positive feedback may be present in allergic processes.

Basophils

Basophils are the other effector cell to express the high affinity IgE receptor. They differ from mast cells by circulating rather than being localised to tissue, and having a shorter life span. They contain similar cytoplasmic granules, and basophil proliferation too is under the control of lymphokines, notably IL-3 and IL- 8. Histamine Releasing Factors are other cytokines produced by T-cells that increase IgE mediated basophil degranulation [O'Hehir et al 1991]. Basophil degranulation releases a number of mediators - histamine, chondroitin sulphate, neutral protease (bradykinin producing), elastase and major basic protein.

Basophil releasability following stimulation with anti-IgE antibody or f-met peptide is increased in atopic individuals compared to nonatopic controls [Casolaro et al 1990]. Furthermore asthmatics exhibit higher responsiveness than allergic rhinitics. Level of basophil releasability is not correlated with mast cell releasability.

Histamine

Undoubtedly, histamine plays a role in a number of allergic diseases, including asthma. It is the commonest agent used in nonspecific bronchoprovocative challenge, where it acts by stimulating H1 receptors on bronchial smooth muscle. However, it should be noted from the outset that the limited effect of antihistamines including modern agents such as terfenadine and ketotifen on asthma suggests that histamine is probably not the main mediator in this disease. Histamine is synthesised and released by mast cells and basophils, but also by other cells, especially those in rapidly growing tissue.

Histamine receptors are ubiquitous throughout the tissues (like histamine - "histos"). Both H1 and H2 receptors are involved in vasodilatation, especially smaller blood vessels, and H1 receptor probably mediate the increase in vascular permeability [Douglas 1980]. Histamine also activates nerve endings, causing release of other mediators such as noradrenaline, eg via axon reflex. These phenomena underlie the classical triple response. General release of histamine causes an anaphylactoid reaction characterised by flushing and hypotension, generalised pruritis (C-fibre stimulation) and urticaria, severe headache, intestinal colic and bronchospasm.

Other immunoglobulins

Abnormalities in immunoglobulins other than IgE are more common among atopic individuals. Particular immunoglobulin G subclass deficiencies (G1 and G3) cause recurrent sino-pulmonary infection, and are overrepresented in asthmatics as well as COPD patients attending specialist clinics [Berger et al 1978; Morgan & Levinsky 1988; O'Keefe et al 1991]. Oxelius [1984] reviewed her own and other studies documenting increased levels of IgG4 in atopic individuals. This subclass may be involved in the development of tolerance. Atopy is doubled in frequency among individuals with IgA deficiency [Strober & Sneller 1991], which often presents in combination with IgG subclass deficiency.

Other substances

Bradykinin

The kinins are oligopeptide mediators with wide ranging effects that include airway smooth muscle stimulation and the regulation of inflammation [Bhoola et al 1992]. They are produced by the action of the kallikreins (a group of serine proteases) on kininogens, and inactivated by kininases (peptidases), the best known of which is kininase II or angiotensin converting enzyme. Kininase Ia has been reported to be increased in intrinsic asthma [Schweisfurth 1989], but not kininase II [Studdy & Bird 1989].

Bradykinin causes bronchoconstriction through several pathways - release of prostaglandins and neuropeptides, C-fibre stimulation and direct effects on smooth muscle. Inhalation precipitates bronchoconstriction in asthmatics, but not in normal subjects. It also stimulates the release of lymphokines, notably IL-1 and TNF, from monocytes, and causes local vasodilation and increased vascular permeability.

Prostaglandins

Both prostaglandins and leukotrienes result from the metabolism of arachidonic acid [Holtzman 1991]. All the prostaglandins are produced from PGH2, which in turn arises from the action of endoplasmic reticulum/nuclear membrane bound enzyme cyclooxygenase. The antagonistic relationship between platelet produced thromboxane A2 (TxA2) and endothelial cell produced prostacyclin (PGI2) is probably the best understood regulatory system involving prostaglandins. TxA2 is a potent vaso- and broncho- constrictor, while PGI2 a vasodilator. PGD2, by contrast is produced by mast cells, and can cause bronchoconstriction and bronchial hyperreactivity, and along with the similarly acting PGF2-alpha, is increased in bronchial lavage fluid from asthmatics [Liu et al 1990]. Both PGD2 and PGE2 potentiate the effects of histamine on vascular permeability, but PGE2 acts as a bronchodilator rather than bronchoconstrictor. Decreased production of PGE2 is one mechanism suggested as the one cause of aspirin hypersensitivity in asthma, since aspirin blocks cyclooxygenase. Another mechanism is a spillover of arachidonic acid into the lipoxygenase pathway with increased production of leukotrienes. PGE2 is known to upregulate T-cell IFN receptor numbers - thus suggesting a negative feedback effect on the allergic response [Holtzman 1991].

In the acute airway response to inhaled allergen, PGD2 is generated and released from mast cells starting shortly after degranulation. Blockade of the prostanoid TP1 receptor attenuates the sustaining of the early phase reaction, once the immediate effects of histamine have subsided [Holgate 1991].

Leukotrienes

There are three lipoxygenases [Sigal & Nadel 1991]. Currently, the most important to asthma is thought to be 5- lipoxygenase that converts arachidonic acid to leukotriene A4 (LTA4). This is then converted to LTB4, LTC4, LTD4, and LTE4. These latter three are the slow reacting substance of anaphylaxis (SRS-A). LTB4 is a eosinophil and neutrophil attractant, while LTC4, LTD4, and LTE4 are potent bronchoconstrictors and bronchial secretogogues. They are produced by a number of inflammatory effector cells; LTC4 by eosinophils, LTC4 and LTD4 by mast cells. These agents are found in lavage fluid from asthmatic airways, while a number of recent clinical studies have shown that lipoxygenase inhibitors and leukotriene receptor antagonists abolish bronchial (and nasal) hyperresponsiveness to challenges such as cold air or exercise in allergic individuals or improve lung function in asthmatics [Hui & Barnes 1991; Impens et al 1993]. For example, Manning et al [1990] showed that a highly specific LTD4 antagonist (MK-571) blocked exercise induced bronchospasm. This would be as efficacious as cromoglycate, another agent that affects the mast cell.

Sigal and Nadel [1991] have highlighted the role of 15- lipoxygenase metabolites in the lung. Tracheal epithelial cells produce large amounts of various forms of 12-hydroxyeicosatetraenoic acid (HETE) in vivo. 8S,15S-HETE is chemotactic to neutrophils, an influx of which might be the first step in an inflammatory cascade though in the guinea pig (ovalbumin) model that Holgate et al [1991] describe, blocking the neutrophil influx preceding the eosinophils in the late asthmatic response does not modify the severity of the reaction. Interestingly, 15-HETE can stimulate LTC4 secretion from mastocytoma cells, and mucus secretion from airway preparations in vitro [Sigal & Nadel 1991].

Lymphokines

As noted earlier when describing the role of the T-cell in IgE production, these molecules have a number of important roles in immunity and inflammation. It is important to remember that these agents synergise as well as cross-inhibit one another, and can have different effects on the same cell at different stages of its lifecycle. The role of IL-12 and IL-13 in atopy has only recently been reported.

IFN-gamma

As reviewed above, gamma interferon inhibits IgE production, but has a vast array of other actions. Its' inhibitory role is ncessary for the development of chronic worm infection

IL-4

IL-5

IL-9

IL-12

IL-13

IL-13 seems to be a key cytokine in the response to gut parasites [Bancroft et al 1998], acting not only to activate immune responses, but also to accelerate gut epithelial cell turnover, leading to worm expulsion [Cliffe et al 2005].

Li et al [1999] showed IL-13 is a potent inducer of eotaxin, and is released by Th2 cells in inflammed airways. Its action is augmented by TNF-alpha, and at least partly mediated via JAK3 (as knockout mice do not develop eosinophilia despite a normal eotaxin levels).

IL-31

Dillon et al [2004] described a novel four-helix bundle cytokine they denoted IL-31 (gene on 12q24.31), largely produced in activated CD4+ T-cells, and especially Th2 cells. The IL-31 receptor is upregulated in IFN-gamma treated monocytes, and in the lung after allergen challenge in the ovalbumin BHR mouse model, and more so in BALB/c mice. A transgenic mouse overexpressing IL-31 develops pruritic alopecia and conjunctivitis, with histopathology reminiscent of atopic dermatitis.

Interferon Regulatory Factors

The IRF (IFN regulatory factor) family (currently) contains four transcription factors that regulate IFN signalling, immune cell development and apoptosis (thus leading to IRF-1's original identification as an oncogene).

IRF-2 seems to have more immune-specific functions (compared to IRF-1) acting to antagonize IRF-1 mediated transcription of IFN and IFN-inducible genes, but independently activating VCAM-1 and Fas-L transcription. It has a specific effect on basophil and megakaryocyte differentiation and proliferation. Elser et al [2002] showed IRF-1 and IRF-2 induced by IFN-gamma bind to IL-4 promoter sites and repress transcription, offering one pathway for cross-regulation.

Chemokines

The chemokines are a set of over 50 cytokines that act as chemoattractants to leukocytes, notably eosinophils in the context of asthma, but also basophils and monocytes. They are usually divided into 4 families, CC, CXC, C, and CX3C, reflecting the disposition of two key cysteines in their amino-acid sequence.

The eotaxins (1,2 and 3), as their names imply, are eosinophil chemoattractants. Eotaxin-1 is increased in the bronchial mucosa of asthmatics, and is correlated with level of BHR.

The CCR3 receptor is the target for the eotaxins, as well as the monocyte chemotactic proteins and RANTES. It is present on eosinophils, basophils, mast cells and (some) Th2 cells. Mouse CCR3 knockouts have altered response to allergen challenge, but this varies according to the protocol. Variable results from knockouts of the other chemokine receptors has been interpreted as being due to redundancy of action [D'Ambrosio et al 2003]. For example, eotaxin-3 has been shown to act as a CCR2 receptor antagonist (repelling monocytes but attracting eosinophils).

Platelet Activating/Aggregating Factor (PAF)

This is an important mediator of inflammation, and its role in asthma is being intensively studied. It derives its name from the original observation that it induces calcium dependent platelet aggregation [Benveniste et al 1972]. PAF (acetyl-glyceryl-ether-phosphorylcholine) is synthesised by Phospholipase A2 from cell membrane phospholipids followed by acetylation by acetyl-(CoA)transferase to the active form. Synthesis is calcium dependent, and the presence of extracellular albumin is required, probably for transport. The half-life of PAF in vivo is only 30 seconds. It is degraded back to its precursor by acetylhydrolase (see below).

Effects of release of PAF

PAF has a number of properties that make it an attractive candidate for involvement in asthma pathogenesis [Smith, 1991]. It is chemotactic to eosinophils, increases airway permeability and mucus secretion. In some (but not all) asthmatic humans, it is a potent bronchoconstrictor of prolonged effect, and Kaye et al [1990] found pretreatment with PAF increased bronchial responsiveness to histamine in six out of eight nonasthmatics. Hsieh [1991] reported that peripheral blood eosinophil and neutrophil counts decreased within five minutes of inhaling PAF in both asthmatic and nonasthmatic children. He interpreted this as being due to margination of these effector cells in the lung (the effect wearing off by ten minutes). Pretreatment with a PAF antagonist (BN52021 - Gingkolide B) blocked the bronchoconstriction due to PAF in all subjects, as well as abolishing the transient leukopenia. In three out of seven asthmatics, it also protected against bronchoconstriction after allergen challenge.

Complement system

Richardson et al [1983] reported that a defect in the alternate complement pathway was found in 10% of 303 healthy infants which persisted to age one year in two-thirds. Frequency was also increased in parents of affected children. Atopy (eczema and/or positive skin test) was increased in these children compared to matched controls (crude OR=6.4), as was respiratory tract infection (OR=4.2). I am unaware of any replications, or of the reliability of the assay used (leucocyte phagocytosis of killed yeast).

Protease Activated Receptors

Protease-activated receptors (PARs) are an unusual G protein-coupled receptors. There are at least four types (1-4). They are activated by cleavage of the amino terminus of the receptor by endogenous proteases, including thrombin and tryptase, and by exogenous proteases, most suggestively the D pteronissinus group 3 and group 9 allergens [Sun et al 2001].

Numbers of PAR-2 are increased in asthmatic lung epithelium [Knight et al 2001]. Activation of PARs leads to smooth muscle hyperplasia, so they are thought to be important in airway remodelling. In the PAR-2 knockout mouse, contact hypersensitivity is attenuated, but not passive cutaneous anaphylaxis [Kawagoe et al 2002].

Sun et al [2001] showed that house dust mite derived proteases cleaved the PAR-2 activation site, and stimulated release of GMCSF and eotaxin from epithelial cells, and effect made refractory by pretreatment with trypsin. Mascia et al [2002] showed D pter extract induced epithelial cells to increased expression of IL-8 and GM-CSF mRNA and protein, an effect blocked by protease inhibitors. They further suggested these mechanisms not as important in skin keratinocytes.

The genes for the PAR family are clustered on chromosome 5q11 (74.9 Mbp from the p telomere, and quite distant from the 5q31 linkage peaks for asthma). The PAR-4 gene (F2RL3) is however on chromosome 19p12.

Nitric oxide and arginase

Nitric oxide (NO) has been recognised as an important local messenger involved in a wide range of inflammation related processes and produced by many effector cell types. Nitric oxide is produced by NO synthase (NOS) from arginine. Many studies have demonstrated increased NO production by lymphocytes and neutrophils in the asthmatic lung.

Another key metabolic pathway for arginine is catalysis by arginase to ornithine and urea. In the macrophase, it is known that arginase I and NOS transcription is reciprocally regulated, and that IL-4 and IL-10 induce arginase production (Modolell et al 1995]. Expression microarray studies of asthma in mouse, guinea pig and monkey have demonstrated arginase to be consistently upregulated in the asthmatic lung. In acute human asthma, Morris et al [2004] found that while exhaled NO levels rose, plasma arginine levels fell, and arginase activity increased.

Xu et al [2003] made consonant findings in mouse experimental autoimmune encephalomyelitis (EAE). These authors also pretreated mice with an arginase inhibitor, which led to a milder EAE phenotype with increased NO production (and decreased IFNG and TNFA production) by splenic mononuclear cells. Chicoine et al [2004] similarly found vascular endothelial cell NO production to be increased by treatment with valine, another arginase inhibitor.

Neurotransmitters

Neurotransmitters play a role in asthma - cholinergic agents such as methacholine and carbachol cause bronchospasm in susceptible individuals, and adrenergic agents reverse bronchospasm. More suggestively, pretreatment with agents such as atropine can block a variety of nonspecific bronchoconstrictors such as cold dry air and inert dust [Kaliner et al 1982].

Szentivanyi [1962, 1968] is usually cited as the first worker to propose the unifying hypothesis that asthma represents an underresponsiveness of the lungs to sympathetic neurotransmitters. He argued that the defect in asthma must be "non-immunological", and suggested the beta adrenergic receptor as the most likely site for this. Specifically, he noted that beta-blockade induced hypersensitivity to histamine and other mediators could be demonstrated in most animal species, as well as in human asthmatics. Nelson [1985] has pointed out that Pottenger suggested the same hypothesis in 1928.

It is now generally accepted that this simple hypothesis as stated is incorrect. Nevertheless, studies continue to be presented that do demonstrate various abnormalities of autonomic function in asthma and in atopy generally. Shah et al [1990], for example, examined 50 asthmatics and found significantly increased responsiveness to vagal/parasympathetic stimulation by carotid sinus massage, Valsalva manoeuvre and deep breath, but not to sympathetic stimulation. The studies reviewed by Kaliner et al [1982] provided evidence for both generalised (multi-system) beta-adrenergic hypo-responsiveness and cholinergic hyperresponsiveness in atopy generally, and additionally alpha-adrenergic hyperresponsiveness in allergic asthma. None of these studies prove that these abnormalities underlie asthma and atopy. Herman et al [1990] suggest the converse. They note that alpha- adrenoceptors on tracheal smooth muscle and on lymphocytes (not present in normal subjects in either case) can be induced in vitro via incubation with histamine.

Noradrenaline

Pharmacological interest in noradrenaline (and adrenaline) and asthma has tended to focus on this agents' effect on airway calibre. Beta-2-adrenoceptors exist in large numbers on bronchial smooth muscle, and upon stimulation lead to muscle relaxation. This smooth muscle is not directly innervated by the sympathetic system however, so adrenergic tone is not important in normal airway regulation. Few alpha-adrenoreceptors are present in the normal or the asthmatic human lung [Spina et al 1989a].

Several authors have demonstrated hyporesponsiveness of airway beta-receptors to beta-agonists in severe asthmatics [eg Spina et al 1989b], but this finding remains controversial, and may reflect treatment with bronchodilators rather than asthma per se. Connolly et al [1992] did find a significant correlation between lymphocyte beta-receptor density and degree of BHR in 26 drug-naive mild asthmatics - these are commonly used as a proxy for airway receptors, which may or may not be valid. A small proportion of untreated or former asthmatics are exquisitely sensitive to beta-blockers, but see below.

Beta-adrenoceptors are also present on airway epithelial cells, submucosal glands [Spina et al 1989b], and mast cells. Sympathomimetics therefore decrease mast cell degranulation and thus allergic wheal responses and allergen-mediated bronchoconstriction [Page 1993].

Paradoxically, Nguyen et al (2009) found that beta-2-adrenoceptor knockout mice can be sensitized to allergen, but have less mucous metaplasia, inflammation and airway hyperresponsiveness. They demonstrated a similar effect using beta-adrenoceptor inverse agonists such as nadolol, but not with "neutral" antagonists such as alprenolol (Penn et al 2009). The proposed mechanism therefore is via proinflammatory effects of beta-2-adrenoceptor stimulation, possibly at lower but chronic levels.

Acetylcholine

Barnes [1990] recently reviewed this important neurotransmitter's functions in the lung. Muscarinic receptors are present in bronchial wall smooth muscle, submucosal glands, and are suspected to be present in human pulmonary vessels. They are more numerous in central airways, numbers falling off in the bronchioles. Stimulation leads to (1) a drop in cAMP levels and thus sympathetically driven processes; (2) activation of phospholipase C, with consequent Ca++ release and bronchoconstriction; (3) activation of phosphokinase C with longer term phosphorylation and uncoupling of adrenoreceptors; and (4) mucus secretion by submucosal glands - by means not currently understood. Sulphur dioxide induced bronchoconstriction is mediated by cholinergic reflex, and can be abolished by pretreatment with the M1 muscarinic receptor blocker pirenzepine. In asthmatics, this is not so for pilocarpine, which targets the M2 receptor, the postganglionic cholinergic nerve (inhibitory) autoreceptor. However, pilocarpine is protective in normal subjects, and so the M2 receptor is another candidate for an underlying defect in asthma. Another interesting observation is that beta-blocker induced asthma is blocked by anticholinergics. Since beta-2-receptors have been found on cholinergic nerves, Barnes has speculated that disinhibited cholinergic overactivity might result from beta-blockade of the M2 deficient asthmatic cholinergic nerve.

Substance P

A number of neuropeptides are thought to be involved in airway submucosal gland secretion. Substance P, for example, causes glandular contraction, and glycoprotein secretion. It is released from C-type nerve fibres. More generally, it acts a bronchoconstrictor, and causes bronchial wall permeability to increase. It is a potent cause of rapid mast cell degranulation (15-20 second timecourse), but does not induce mast cell synthesis of prostaglandins and leukotrienes [Church et al 1991].

Neurokinins

Neurokinin A is a more potent bronchoconstrictor than Substance P, though less of a secretogogue. Since neutral endopeptidase (the enzyme that degrades neurokinins) is produced by airway epithelium, epithelial damage, as is seen in asthma, potentiates Neurokinin A's action.

Vasoactive Intestinal Peptide (VIP) is present in both airways, pulmonary and bronchial arteries, and within a number of effector cells implicated in asthma, such as eosinophils, neutrophils, mononuclear and mast cells. In the latter it is released during degranulation. VIP causes bronchial smooth muscle relaxation - it is 100 times more potent than isoproterenol as a bronchodilator [Barnes 1991] - but also inhibits T-cell, mononuclear and mast cell function, as well as platelet aggregation. In its latter role it is a potent anti-inflammatory. However, there is as yet no evidence of a (presumable) deficit of VIP in asthma. It too leads to increased submucosal gland secretion.

Alpha-1-antitrypsin deficiency (AAT)

Complete deficiencies of this circulating (serine) protease inhibitor were first noted to be associated with COPD by Eriksson [1964]. A similar association was pointed out for AAT partial deficiencies by Lieberman [1969]. Interestingly, a number of studies have suggested both that the diagnosis of asthma is more frequent among relatives of probands with the complete forms of AAT deficiency, and that asthmatics are more likely to suffer from intermediate deficiencies.

Alpha-1-antitrypsin is a 52 kD globular protein and is coded for by an codominant gene (symbol PI) located on chromosome 14q31.2 [Kalsheker & Morgan 1990]. The common normal allele is M; those associated with deficiency states are the S (E264V) and Z (E342K) alleles. The gene frequency of the Z allele is approximately 2% in European populations. A number of less common variants have also been identified. The PiZZ phenotype is the most severely affected (by panacinar emphysema and cirrhosis), while the PiSZ and PiSS phenotypes are at increased risk of lung disease, especially following exposures such as cigarette smoking. The disease association with PiMS and PiMZ is smaller and is not seen in all studies.

Fagerhol and Hauge [1969] have reported high proportions of PiMS and PiSS subjects in asthma patients while Hyde et al [1979] also found that PiMS and MZ asthmatic children required more intense drug treatment than control PiMM asthmatics. Buist et al [1979] found an excess of asthma in heterozygotes in a matched 2:1 case-control study (3/21 cases and 2/42 controls). The strength of this design was that probands were selected as parents of homozygotes detected through a screening program. Mano et al [1975] found similar patterns in 127 Japanese asthmatics (5.4% were carrying variants) versus 965 blood donors (3.2%).

Townley et al [1990] described results of AAT phenotyping of 723 subjects from the Natural History of Asthma study. These included the families of asthmatic probands, and families ascertained for a three generation absence of atopic disease. All completed a questionnaire, underwent SPT, total sIgE level estimation and methacholine provocation. Asthmatics made up 36% of the PiMS group and 21% of the PiMM group (P=0.04). It is not surprising to find therefore that the PiMS group also had a significantly lower mean provocative methacholine dose (for PiMS: Area35=1417; for PiMM 2617). However, removing the current and former asthmatics did not alter this difference. Serum IgE level was the same for all three phenotypes.

Gaillard et al [1992] reported similar findings for subgroupings of the MM phenotype. The M2M2 phenotype was found to be more frequent in 90 asthmatics than 240 controls. Asthmatics were also found to have higher plasma levels of alpha-1-AT, but lower levels of Elastase Inhibitory Capacity (EIC). These findings are suggestive, especially when one notes that EIC level and the EIC/alpha-1-AT level ratio (a measure of molar efficiency of the different Pi types) are lowest for PiM2M2 of the MM subtypes, with the exception of the M2M3 subtype [Oakeshott et al 1985].

In Nigerian asthmatics, Awotedu and Adelaja [1990] found 74/99 asthmatics to be MM phenotype compared to 98/100 controls. The MZ phenotype was present in 19 asthmatics and one control. In a similar study of Puerto Rican asthmatics [Colp et al 1990] in New York found 41/55 nonsmoking asthmatics to be MM, and 49/61 controls - not a significant difference.

Eden et al [2003] report on asthma-like symptoms in a cohort of 1052 PiZZ patients. Two-thirds had experienced more than one episode of wheezing, and 90% had wheezed with a cold. A positive bronchodilator test (12% improvement) was present in 49%, and a total sIgE>100 IU/ml in 17%. Those with an elevated IgE were not more likely to exhibit a positive bronchodilator response. They did experience a faster rate of FEV1 decline with age, but this was not significant after adjusting for bronchodilator response, age, sex, airflow obstruction and smoking.

Alpha-1-antichymotrypsin deficiency

Lindmark has performed a number of studies of alpha-1-ACT, deficiencies of which lead to a clinical syndrome similar to that associated with alpha-1-AT deficiency [Eriksson et al 1986]. In a recent study [1990], Lindmark screened twelve women heterozygous for alpha-1-ACT and their relatives for asthma and allergic rhinitis, comparing them to a group of 58 matched controls. The index cases were three times more likely to report asthma (CI for OR 1.05-9.80), but not hayfever. Relatives with decreased levels of alpha-1-ACT (N=15) were also more likely to report asthma (OR=3.1; 95% CI=0.96-9.83, P=0.06). Since this deficiency occurs in only 0.5-1% of the Swedish population, it cannot be a major determinant of asthma.

Physiological determinants of airway calibre

BR, symptoms and changes in airway calibre

Although I have noted that bronchial hyperresponsiveness is usually accepted to be necessarily present in asthma, the relationship between severity of non-specific bronchial hyperresponsiveness and presence or severity of asthma symptoms is not perfect [Josephs et al 1990; Pattemore et al 1990]. This must be due partly to technical factors in the determination of bronchial response (see Chapter 4), biological variation in BR (for example, circadian changes have been reported), and variation in the relationship between symptomatic disease and BHR.

Geometrical considerations

Macklem [1990] has reviewed his own and others' work to provide a unifying physical model of the determination of airway calibre. He first notes that normal individuals can inhale very large doses of methacholine and other "physiological" bronchoconstrictors without airway resistance falling past a "plateau" value. He argues that this represents the maximal level of airway closure bronchial smooth muscle contraction can achieve unaided. This is because increasing contraction stretches the supporting alveoli leading to an opposing elastic recoil. Evidence for this is the effect of lung volume on airway resistance, in that the maximal airway resistance in normal subjects is greatest at or below FRC, and decreases rapidly above FRC. Further narrowing of the airway can only occur if the submucosa is thicker than normal due to cellular infiltration and oedema - known to be present to some extent even in asymptomatic asthmatics. Alternatively, the peribronchial adventitae may be oedematous, attenuating the effects of elastic recoil. It should be noted that a number of recent biopsy studies have shown that the bronchial walls of asthmatics have decreased elastic layers and subepithelial fibrosis. This might also explain a fall in elastic recoil. Macklem [1991] describes this as an uncoupling or unlinking of airway from parenchyma. Wiggs et al [1992] using these assumptions in a mathematical model of the airway produced simulated dose-response curves for bronchoprovocative challenge that fit well to observed curves for asthma and COPD.

Intraluminal mucus

Airway mucus, especially in a previously narrowed airway, leads to obstruction of the lumen and to wheeze. Mucus plugging is a characteristic feature of status asthmaticus, and partly explains the limited response in this condition to bronchodilators.

Respiratory epithelium and submucosa

A number of studies have documented submucosal fibrosis in asthmatic airways [Roche et al 1989]. This is located just below the basement membrane, and is comprised of fibronectin, collagen III and V. Its deposition is thought to follow mucosal inflammation. A smaller baseline bronchial calibre has been demonstrated in small airway disease.

Bronchial wall smooth muscle

Bronchial smooth muscle contraction is probably the most important mechanism involved in acute bronchoconstriction. Earlier theories have looked for hypersensitivity or hypercontractibility in smooth muscle as the sole defect in asthma. Holgate and co-workers have presented evidence that during chronic airway inflammation, infiltration by myofibroblasts occurs, leading to the increase in depth of smooth muscle in asthmatic bronchi/ioles. These cannot be distinguished from myoblasts on conventional staining.

Peribronchial adventitia and vessels

It has been increasingly obvious that the state of the airway circulation has effects on airway calibre either by (1) increase in vessel wall permeability due to inflammatory mediators, (2) changes in clearance of such mediators, or (3) a direct influence of increased bulk of the dilated vessel. The first mechanism is self-explanatory. Dinh Xuan [1990] invokes the second as an explanation for the protective effects of prostacyclin given prior to methacholine challenge and the third mechanism as more important as an explanation for the effects of methoxamine (an alpha-1 adrenergic agent) in preventing exercise induced asthma, blocking the post-exercise rebound hyperemia that may cause this. Direct measurement of tracheal mucosal height in response to vasoactive agents, and the bronchoconstriction brought on by intravenous saline load are two lines of evidence further supporting this third mechanism.