The immunology of virus infection in asthma

Infection of the respiratory tract by viruses is common and can result in a variety of specific syndromes such as the common cold, pharyngitis, tracheobronchitis, croup, bronchiolitis or pneumonia. Such syndromes may be superimposed on a background of pre-existing chronic respiratory disease. One such chronic respiratory disease is asthma. Numerous virus types can cause respiratory disease. Infection may be localized to the respiratory tract, as in respiratory syncytial virus (RSV) infant bronchiolitis, or part of a generalized systemic illness such as measles or chickenpox. Each viral pathogen can produce more than one clinical syndrome and each syndrome may be caused by a range of different viruses (table 1⇓). The pathology resulting from virus infection is influenced by host factors including age, previous infection or immunization, pre-existing respiratory or systemic disease and immunosuppression/compromise. The nature and severity of disease observed is dependent both on the direct harmful effects of the virus itself and on damage caused to host tissues as a consequence of the host immune response to the virus. Some immunopathology may be unavoidable if the host is to eradicate the virus. An ideal immune response would result in early elimination of the virus with minimum harm to the host.

Asthma is a disease of major importance affecting 20–33% of children in the UK 1. The health costs of this condition are enormous in terms of absence from work or school, general practise consultations, hospital admissions and mortality. It is a multifaceted syndrome involving atopy, bronchial hyperreactivity and immunoglobulin (Ig)-E and non-IgE-mediated acute and chronic immune responses. The asthmatic airway is characterized by an infiltrate of eosinophils and of T-lymphocytes expressing the type 2 cytokines, interleukin (IL)-4, IL-5 and IL-13. Trigger factors associated with acute exacerbations of asthma include exposure to environmental allergens, especially animals, moulds, pollens and mites, cold, exercise and drugs. The link between respiratory infection and asthma exacerbation is well established, although incompletely understood. In the 1950s, this association was attributed to bacterial allergy 2, but it is now clear that the majority of exacerbations are due to viral rather than bacterial infection.

The antiviral immune response

Current concepts of a typical antiviral immune response, as reviewed in detail elsewhere 3, 4, result from a huge body of research carried out using human volunteers, patients and experimental animals, especially inbred mice. The results of animal studies may not be directly applicable to the outbred human population, but are valuable because ethical considerations often limit direct investigation of the human immune system. A typical response involves a combination of nonspecific (innate) and specific immunity.

Nonspecific elements include: phagocytes (e.g. neutrophils and macrophages) that engulf and destroy viruses; natural killer (NK) cells that recognize and destroy virus-infected cells on the basis of alterations to normal cell surface proteins; cells including NK cells, neutrophils, macrophages, mast cells, basophils and epithelial cells that release cytokines (e.g. the interferons), which have immunoregulatory or antiviral actions; and components of body fluids that are capable of neutralizing viral infections independently, or in combination with antibodies.

Specific immunity involves the production of antibodies by B-lymphocytes and the activities of cytotoxic T-cells following the processing and presentation of viral antigens by additional cells of the immune system, the most important of which are probably the dendritic cells. An important feature of specific immunity is memory, which modifies the overall response to re-infection by a previously encountered respiratory virus and alters the timing and magnitude of contributions due to the different components.

In a primary infection, the virus multiplies in the respiratory tract, reaching peak levels at ∼2 days. Type I interferons (IFNs) are first detected at this time, peaking at day 3 and falling to an undetectable level by day 8. IFNs activate NK cells, which are first detectable at day 3 and then peak at day 4. In addition to the destruction of virally-infected cells, NK cells release cytokines, including IFN-γ that activate additional inflammatory cells in the airway. Such nonspecific immune mechanisms are essential for early defence against the virus during the first few days after infection. At this early stage of the antiviral immune response, viral antigens are processed locally in regional lymph nodes by dendritic cells and they are then presented to T-cells. CD4+ and CD8+ T-cells are detectable at day 4 and day 6. CD8+ cytotoxic T-cell responses peak at day 7, then generally decline, becoming undetectable by day 14. However, memory CD4+ and CD8+ responses may persist for life. T-cell recruitment is dependent both on the production of chemokines and on alterations in the expression of adhesion molecules on the endothelium of inflamed tissues. Time is also required to generate B-cell responses. Mucosal IgA may be detected at day 3, serum IgM at day 5–6 and IgG at day 7–8; they all then increase in amount and avidity for a period of 2–3 weeks. IgA falls to undetectable levels after 3–6 months, whereas serum IgG may remain detectable for life. Specific immune mechanisms such as CD8+ T-cells and Ig are responsible for the eradication of the infectious virus usually within 7 days after infection.

A secondary infection with the same virus results in rapid mobilization of B- and T-cell-specific immunity, with an earlier T-cell peak coinciding with the NK cell peak at day 3–4. If re-infection is with the same serotype, a rapid increase in levels of pre-existing neutralizing antibodies may limit viral replication to such an extent that infection is clinically silent. As fewer cells are infected, there is relatively less activation of nonspecific immunity and it may be difficult to detect a CD8+ T-cell response.

Is this classical immune response to respiratory virus infection modified in the context of asthma? This question is the subject of the present review.

Epidemiology

Viral respiratory tract infections are a major cause of wheezing in infants and adult patients with asthma. Their role may have been underestimated in early epidemiological studies because of difficulties with isolation and identification 5. The introduction of molecular biological techniques (e.g. the polymerase chain reaction (PCR), in situ hybridization, reverse transcriptase (RT)-PCR and in situ PCR) to such studies has implicated viral infection in the majority of asthma exacerbations.

Indirect evidence from population studies has established a significant correlation between the seasonal variation in wheezing episodes in young children and peaks of virus identification 6. Seasonal patterns of identification of respiratory viruses are associated with peaks in hospital admissions for both children and adults with asthma, indicating a role for such infections in severe asthma attacks 7. Studies that showed an increased rate of virus detection in individuals suffering from asthma attacks have provided direct evidence implicating viral infection in asthma exacerbations. Viruses have been isolated in 10–85% of asthma exacerbations in children 6, 8–10 and in 10–44% in adults 11, 12. The highest rates of identification were found in: 1) studies where subjects were followed prospectively, allowing collection of clinical specimens early during the course of the illness; 2) where PCR-based methods of diagnosis were used in addition to serology and culture; and 3) where the methodology used allowed for detection of rhinoviruses (RV). The rate of detection of viruses between exacerbations, when individuals are asymptomatic, is only ∼3–12%. In contrast, a study of transtracheal aspirates in adult asthmatics during exacerbations yielded sparse bacterial cultures, with no correlation to clinical illness and no difference from those of normal subjects 13.

In almost all studies of asthmatics the predominant viruses are RVs, RSVs and parainfluenza viruses. RV alone is detected in ∼50% of virus-induced asthma attacks. Adenoviruses, enteroviruses and coronaviruses are also detected, but less frequently. Influenza is only found during annual epidemics.

Evidence for a role for viruses in exacerbations of other chronic respiratory diseases is more limited. Recent work suggests that virus infection, RV in particular, is important in chronic obstructive pulmonary disease and that viral infections may contribute to the decline in lung function that is observed over time in the patient group studied 14, 15.

Experimental virus infection

Studies of experimental respiratory virus infection in human volunteers are limited by concerns of safety 16. Many of these studies have, therefore, focused on the experimental innoculation of RV in allergic rhinitis, mild asthmatics or normal control subjects 17–29. Such studies provide a useful model of natural virus infection in asthma and offer advantages including: 1) patient selection and monitoring, under controlled conditions, before, during and after infection; and 2) observation of RV-induced effects, including asthma symptomatology, changes in the use of medication, lung function and airway pathology/immunology.

In general, the clinical, physiological and cellular responses to experimental RV infection in asthma are relatively mild and do not necessarily mimic exactly the events after a natural common cold. It has been suggested that this requires a more complex model including both virus infection and pre-existing increase in allergic airway inflammation. Indeed, recent epidemiological evidence confirms a synergistic interaction between virus infection and allergen exposure in precipitating hospital admissions for asthma. Most studies of experimental virus infection in allergic subjects are performed outside the relevant season for allergen exposure. One recent attempt to provide such a model utilized RV infection in subjects with allergic rhinitis. Individuals received three high-dose allergen challenges in the week prior to innoculation to mimic combined allergen exposure and virus infection 30. Interestingly, prior allergen challenge in this model, somewhat unexpectedly, appeared to protect against a RV cold with delayed nasal leukocytosis, increased generation of the pro-inflammatory cytokines IL-6 and IL-8 and a delayed, less severe clinical course. There was an inverse correlation between nasal lavage eosinophilia and the severity of cold symptoms. The authors of this study propose that limited high-dose allergen challenge may not reproduce the effects of chronic low-dose allergen exposure and may stimulate the production of anti-inflammatory mediators such as IL-10 or antiviral cytokines such as IFN-γ or tumour necrosis factor (TNF)-α. Further development of models of combined allergen exposure and virus infection is clearly required.

Rhinovirus infection of the lower airway

While other respiratory viruses (e.g. influenza, parinfluenza, RSV, adenovirus) are well-recognized causes of lower airway syndromes such as pneumonia and bronchiolitis and are capable of replication in the lower airway, until recently the experimental RV infection model did not provide clear evidence to show that RV infection occurred in the lower airway as well as in the upper respiratory tract. Although the possibility of nasopharyngeal contamination could not be ruled out, RV has been detected in lower airway clinical specimens such as induced sputum 31, tracheal brushings 26 and bronchoalveolar lavage (BAL) 32 both by (RT)-PCR and culture, but for all such specimens, contamination from the nasopharynx could not be excluded. RV has been cultured in cell lines of bronchial epithelial cell origin 33 and replication has been demonstrated in primary cultures of bronchial epithelial cells 34, 35. The preference of RV for culture at 33°C rather than 37°C has been used as an argument against lower airway infection, but now there is evidence that replication does occur at lower airway temperatures 36. Finally the use of in situ hybridization has demonstrated RV in bronchial biopsies of subjects following experimental infection 34. These data confirm that RV infection of the lower airway does occur and they directly implicate lower airway infection in the pathogenesis of asthma exacerbations.

Physiological effects of experimental rhinovirus infection

Subjects with asthma and/or allergic rhinitis exhibit increased pathophysiological effects as a result of RV infection as compared to nonatopic, nonasthmatic controls. With detailed monitoring, it is possible to detect reductions both in peak flow 37 and home recordings of forced expiratory volume in one second (FEV₁) in atopic asthmatic patients in the acute phase of experimental RV16 infection 24. There is an enhanced sensitivity to histamine and allergen challenge after RV16 inoculation in nonasthmatic atopic rhinitic subjects 19, 29. RV16 increases asthma symptoms, coinciding with an increase in the maximal bronchoconstrictive response to methacholine ⩽15days after infection 20. There is a significant increase in sensitivity to histamine in asthmatic subjects after RV16 infection, most pronounced in those with severe cold symptoms 25.

Interactions between virus infection and asthmatic airway inflammation

The interaction of respiratory virus infection and chronic asthmatic airway inflammation results in respiratory symptoms that are more severe than those suffered by nonasthmatic individuals. The detailed immunological mechanisms underlying this interaction are currently unclear. The disease syndrome following viral infection is a consequence both of the direct harmful effects of the virus and of immunopathology resulting from the host immune response. In an asthmatic individual, exacerbation may occur because of the functional interaction between viral pathology and asthmatic pathology (i.e. through different mechanisms with the same end effect on function or by sharing the same pathogenetic mechanism in an additive or even in a synergistic fashion). Pre-existing asthmatic inflammation might interfere with an effective antiviral response, thus allowing the virus itself to cause increased airway damage. Alternatively, virus infection might increase the sensitivity of the asthmatic airway to trigger factors such as allergen exposure. In fact, it is likely that virus-induced asthma exacerbations occur because of a combination of these four types of interaction. The increased severity of symptoms (including lower respiratory symptoms), seen in nonasthmatic subjects with allergic rhinitis during viral infection suggests that the atopic phenotype itself is important in determining the clinical syndrome following infection by respiratory viruses. However, it is possible that virus infection, in some way, amplifies subclinical allergic lower airway inflammation already present prior to infection.

Table 2⇓ summarizes some of the current hypotheses proposed to explain the mechanisms of exacerbation of asthma following respiratory virus infection. The evidence supporting these hypotheses is reviewed in detail later.

Effects of viruses on airway epithelial cells

Respiratory viruses enter into and replicate within epithelial cells lining the lower airways. Entry is dependent on interaction with specific receptors, for example intercellular adhesion molecule (ICAM)-1 in the case of the major group RVs and the low density lipoprotein receptor in the case of the minor group RVs. Influenza viruses bind sialic acid residues via haemaglutinin.

The upregulation of ICAM-1 in the asthmatic airway is one explanation for the increased severity of RV infection. RV itself has been shown to further upregulate ICAM-1 in bronchial biopsies, following experimental RV infection 38. In nasal epithelial cells obtained by brushings from atopic subjects, basal levels of ICAM-1 were increased relative to nonatopic subjects and were elevated in the relevant season for peak allergen exposure. Nasal epithelial cells from atopic subjects showed further upregulation after in vitro culture with allergen. The highest basal level of expression of ICAM-1 was found on nasal polyp epithelial cells and this was increased further after infection with RV14. Viral titres recovered after RV14 infection were significantly higher for polyp epithelial cells than for nonatopic and atopic nonpolyp epithelial cells 39.

In vitro RV increases expression of both ICAM-1 and vascular cell adhesion molecule-1 in cultures of primary bronchial epithelial cells and in the A549 respiratory epithelial cell line, via a mechanism involving the transcription factor nuclear factor (NF)-κB 40, 41. Inhibition of the upregulation of ICAM-1 might be expected to improve the course of RV infection. One effect of corticosteroids is to inhibit NF-κB 42. In A549 cells as well as primary bronchial epithelial cells, pretreatment with three different corticosteroids, (i.e. hydrocortisone, dexamethasone and mometasone furoate), inhibits RV16-induced increases in ICAM-1 surface expression, messenger ribonucleic acid (mRNA) and promoter activation, without alteration of virus infectivity or replication 43. Disappointingly, a study of inhaled corticosteroids in asthmatics prior to experimental RV infection failed to show a reduction of virus-induced ICAM-1 expression in bronchial biopsies 38. It is possible, however, that a longer course and/or a higher dose of inhaled steroid, or administration of oral steroids, might have demonstrated a significant effect.

The extent of epithelial cell destruction observed in the airway varies according to virus type. Influenza typically causes extensive necrosis 44, whereas RV causes little or only patchy damage. Destruction of epithelial cells results in an increase in epithelial permeability, increased penetration of irritants and allergens and exposure of the extensive network of afferent nerve fibres. These effects may contribute to increased bronchial hyperresponsiveness.

There is increasing evidence that the epithelium does not simply act as a physical barrier, but has important regulatory roles. Epithelial cells contribute to the immune response following virus infection by producing cytokines and chemokines (fig. 1⇓). They may also act as antigen-presenting cells, particularly during secondary respiratory viral infections. Furthermore, epithelial cells express major histocompatibility complex (MHC) class I and the co-stimulatory molecules B7-1 and B7-2, and this expression is upregulated in vitro by RV16 45.

Inflammation is a central event both in asthma and viral infection. The processes involved include interacting cascades from the complement, coagulation, fibrinolytic and kinin systems of the plasma as well as cell-derived cytokines, chemokines and arachidonic acid metabolites. The understanding of the interaction of viruses with these cascades in asthma is incomplete, and it is likely that different viruses interact with each system to different extents. However, it is reasonable to believe that in all cases the initial trigger of the inflammatory reactions is an epithelial cell-virus interaction.

A multitude of inflammatory mediators are generated or act on the epithelial surface. Bradykinin, a polypeptide consisting of nine amino acids, is generated from plasma precursors as part of the inflammatory process and has been shown to be present in the nasal secretions of RV-infected individuals 46. Given intranasally, bradykinin is able to reproduce some of the symptoms of the common cold, such as sore throat and rhinitis 47. Although the presence of kinins in the lungs of virus-infected individuals has not been reported, they are present both in the upper and lower airways in allergic reactions 48–50.

Some viruses may also cause complement-mediated damage. Complement components (C) bind to epithelial cells both in vitro and in vivo during RSV infections. C3a and C5a are increased in human volunteers infected with influenza A virus 51.

Nitric oxide (NO) is produced by diverse sources including epithelial, endothelial and smooth muscle cells. In human airways, NO appears to be important in relaxation of the human airway smooth muscle 52. In experimental animals, parainfluenza virus-induced hyperreactivity correlates with a deficiency in constitutive NO production 53. On the other hand, in inflamed tissues, NO reacts with superoxide anion generating peroxynitrite, a highly toxic compound, suggesting a dual role for this mediator. Increased levels of exhaled NO are found in nonasthmatic volunteers following natural colds 54 as well as in asthmatic patients after experimental RV infection 55. In the latter study, an inverse association between NO increase and deterioration of airway hyperresponsiveness was demonstrated, arguing in favour of a protective role for this substance. This is supported further by the observation that NO reduces cytokine production and viral replication in an in vitro model of RV infection 56.

Viral infection of the respiratory tract results in significant changes in the pattern of cytokine expression by a number of cell types. These include cells of the immune system, which may be increased in number and activation status, and other types of cells (e.g. epithelial cells) often considered to be structural, but which in fact contribute significantly to the immune response. Efficient orchestration of the immune response by cytokines is essential for eradication of the virus. Modification of cytokine expression in the airway may contribute to the increased severity of virus infection in asthma.

In vitro studies of bronchial epithelial cell lines have demonstrated the production of a wide range of pro-inflammatory cytokines (e.g. IL-1, IL-6, IL-11, IFN-α, IFN-γ, TNF-α and granulocyte macrophage colony stimulating factor (GM-CSF)) and the chemokines (e.g. IL-8, regulated on activation, normal T-cell expressed and secreted (RANTES) and macrophage inflammatory protein (MIP)-1α) in response to RV and RSV 33, 57, 58. In vivo, these cytokines can be found in nasal lavage in association with RV infection 59.

The specific roles of individual cytokines in the human lower airway during viral infection are not well understood, but increasing information is becoming available. IL-1, TNF-α and IL-6 share pro-inflammatory properties such as the induction of the acute phase response and the activation of both T- and B-lymphocytes. IL-1 enhances the adhesion of inflammatory cells to endothelium, facilitating chemotaxis 60. TNF-α is a potent antiviral cytokine, but in vitro it increases the susceptibility of cultured epithelial cells to infection by the major group RV, RV14, through upregulation of ICAM-1 33. IL-6 has been shown to stimulate IgA-mediated immune responses. IL-11 may also be important in virus-induced asthma 61. It appears to cause bronchoconstriction by a direct effect on bronchial smooth muscle 58. Production of this cytokine by human stromal cells in vitro is increased by RV14, RSV and parainfluenza type 3, but not by cytomegalovirus, herpes simplex virus-2 or adenovirus. In vivo, IL-11 is elevated in nasal aspirates from children with colds; levels correlate with the presence of wheezing. Similarly, the chemokine MIP-1α is increased in nasal secretions during natural viral exacerbations of asthma 62. Studies in MIP-1α knockout mice suggest that it mediates pneumonitis as a result of influenza 63.

Viral upregulation of cytokines may be mediated through certain key transcription factors. Increases in IL-6 and IL-8 production by cultured epithelial cells due to RV9 are dependent on NF-κB 41, 64 as is the induction of IL-1, IL-6, IL-8, IL-11 and TNF-α by RSV 65, 66.

Effects of viruses on airway smooth muscle cells

Studies utilizing isolated rabbit tissues and human cultured airway smooth muscle cells suggest that RV16 exposure may have a direct effect on smooth muscle cells, thus resulting in increased contractility to acetylcholine and impaired relaxation to isoprotenol. This effect is dependent on ICAM-1 and appears to involve an autocrine signalling mechanism including upregulation of the production of IL-5 and IL-1β by the airway smooth muscle itself 67. Whether RV reaches airway smooth muscle cells in sufficient quantity to produce a significant effect by this mechanism in vivo is as yet unknown. The effects of other respiratory viruses on smooth muscle require further investigation.

The cellular immune response to virus infection in the lower airway

A variety of leukocytes show changes in number, site of accumulation and activation state in response to virus infection. Since these cells are also implicated in asthmatic inflammation of the lower airways, they provide potential sites of interaction between the immunopathologies of virus infection and asthma.

The effects of pre-existing immunity on virus-induced exacerbations of asthma

The effect of pre-existing immunity on virus-induced exacerbation of asthma is an area that requires further investigation.

The presence of pre-existing serum-neutralizing antibodies to RV39 and to RV-Hanks modify the clinical course of experimental infection in human subjects 103, 104. Local IgA and IgG, passing from the vasculature into the pulmonary interstitium, contribute to viral clearance. There are, however, >100 recognized serotypes of RV, thus making repeated infection with the rhinoviruses to which an individual lacks appropriate antibodies a common experience.

T-cell responses to RV demonstrate MHC class I restricted cross-reactivity between serotypes due to specificity for conserved epitopes within the viral capsid proteins 1–3 105. RV16- and RV49-specific T-cell clones from human peripheral blood demonstrate recognition of both serotype-specific and shared-viral epitopes 106. Vigorous proliferation of, and IFN-γ production by, PBMC in response to RV16 in seronegative subjects is associated with reduced viral shedding after innoculation 107. As a consequence of serotype cross-reactivity, recurrent infection with RV of different serotypes would be expected to modify T-cell responses to subsequent infection, either enhancing the antiviral response or, in the case of asthmatic subjects, further amplifying allergic inflammation.

Immunization against RV is not currently practical because of the large number of serotypes, and in the case of other respiratory viruses, immunization can have unexpected consequences. Vaccination of susceptible individuals against prevalent strains of influenza provides protection against disease. Unfortunately, infant trials of a formalin-inactivated, alum-adjuvanted vaccine for RSV were less successful, resulting in a high proportion of vaccinated infants developing lower respiratory tract disease (similar to severe bronchiolitis of infancy) when infected with RSV later in childhood. The vaccine induced complement-fixing and neutralizing RSV antibodies, but appeared to sensitize the vaccinees to develop significant immunopathology upon natural infection with the virus.

Future directions

Respiratory viruses are important triggers of wheezing illness or asthma. RV is common in all age groups, and RSV is most important in infants and young children. The mechanisms involved in virus-induced asthma exacerbations are complex and incompletely understood. The pathology observed in the airway is dependent on virus type. RSV and influenza are capable of causing extensive epithelial necrosis, whilst RV generally produces less destruction. Overall, the disease syndrome suffered by the host results from a combination of direct viral effects on lung tissues and of immunopathology associated with the immune response to the virus, some of which may be unavoidable if the virus is to be eliminated. The virally-infected epithelial cell is an important component of the antiviral immune response, producing cytokines and chemokines capable of activating and recruiting a variety of other cells including lymphocytes, eosinophils and neutrophils. Efficient clearance of a virus is orchestrated by antibodies and by T-cells producing type 1 cytokines. The asthmatic airway is rich in type 2 cytokines and this may result in virus-specific T-cells with type 2 or mixed type 1/type 2 character. If this is the case, then virus infection could be followed both by an inefficient antiviral immune response, with delayed viral clearance, and by amplification of ongoing asthmatic inflammation; the consequence of this interaction is severe, often causing prolonged viral illness and exacerbation of asthma (fig. 2⇓).

Current treatment for virus-induced asthma exacerbations is limited to high-dose inhaled and/or oral corticosteroids, which are only partially effective and invariably associated with side effects, and to purely symptomatic treatment with bronchodilators. Antiviral therapy exists for influenza; however, it is not available for the most common respiratory viruses, RV and RSV. Vaccination is difficult for RV because of the many serotypes, and caution surrounds the development of vaccines for RSV, which must be designed to avoid subsequent enhanced immunopathology.

As an alternative approach, virus-induced inflammation could be treated by strategies that promote type 1 responses in individuals with excessive type 2 responses. Further work is needed to elucidate the important sites of interaction between the immunological networks of asthma and of virus infection. Greater knowledge is required if the key targets for therapeutic intervention are to be identified, the aim of which will be to minimize immunopathology, whilst maintaining or enhancing the host antiviral immune response.