INTRODUCTION
It is increasingly appreciated that symptoms and signs of many viral diseases are caused less by viral cytopathic effects than by the host's response to infection. The peak of viral infection often precedes the period of maximal illness, which coincides with cellular infiltration of infected tissues and the release of inflammatory mediators. In this review, we discuss the role of overexuberant immune responses in disease caused by respiratory syncytial virus (RSV).
RSV is the most important cause of viral respiratory tract infection in infants. Previous reviews have described the clinical impact of RSV disease (
63), its pathogenesis (
102), and the molecular biology of paramyxoviruses (
30) and have compared RSV to other paramyxoviruses (
40). The aim of this review is to provide an up-to-date summary of the host-RSV interaction and how this can cause disease. We will discuss the burden of disease caused by RSV infection, factors which affect disease severity, and what is known of the mechanisms of viral bronchiolitis.
It is useful to consider inflammation in RSV disease in three distinct scenarios: (i) the response to first infections in previously nonexposed hosts, (ii) the pathogenesis of enhanced disease in RSV-infected recipients of formalin-inactivated RSV (FI-RSV) vaccines, and (iii) specific animal models of disease augmentation. By comparing and contrasting the immunopathogeneses of primary bronchiolitis and enhanced disease, we attempt to identify common mechanisms that are shared or distinct in these conditions.
VIRAL INTERACTIONS AFFECTING DISEASE PATHOGENESIS: THE INNATE IMMUNE RESPONSE AND VIRAL EFFECTS ON THE HOST
RSV surface proteins bind glycosaminoglycans (e.g., heparin or chondroitin sulfate), removal of which reduces the infectibility of HEp2 cells in vitro (
65). RSV can also interact with annexin II and L-selectin (
99). The RSV glycoprotein G has been shown to have structural homologies with the CX3C chemokine fractalkine. G binds the human CX3CR1 receptor and mediates chemotaxis of cells that respond to CX3CL (
169). It is possible that this interaction facilitates binding to CX3CR1-bearing cells, including mast cells and neuronal cells. The fusion (F) protein binds TLR4 (
89), upregulating its surface expression and sensitizing airway epithelial cells to endotoxin (
109). The frequency of TLR4
+ monocytes is increased in the peripheral blood of some infants with RSV bronchiolitis (
46), but the role of TLR4 in vivo is unclear (
41).
Events during the first minutes and hours after viral entry are of key importance, not only in determining the balance between viral multiplication and elimination but also in setting the pattern that will be followed by acquired immune responses. Early viral proteins therefore frequently interfere with innate immune mechanisms (Fig.
1).
Once within cells, RSV upregulates the STAT pathway via reactive oxygen species (
95). Nitric oxide production is associated with the upregulation of IL-8 (polymorphisms in the promoter of which have been shown to be associated with bronchiolitis severity [
54]), leading to pulmonary neutrophilia (
159). RSV infection also upregulates proapoptotic factors in the cell (
88) and activates the nuclear factor κB (NF-κB) pathway (
19), which stimulates the transcription of genes directly involved in the antiviral response via IκB kinase (
60). NF-κB is an upstream mediator of many of the innate responses, especially alpha/beta interferon and chemokine production, which leads to the recruitment and activation of cells and the production of further inflammatory mediators. RSV's nonstructural proteins, NS1 and NS2, cause species-specific resistance to alpha/beta interferons (
141,
157) via interferon regulatory factor 3 (IRF3) (
21,
22). Similar effects have been demonstrated with other paramyxoviruses, simian virus 5, and Sendai virus (
38).
Chemokines are crucial in directing the recruitment of different cell subsets and make attractive targets for intervention. Double-stranded RNA selectively induces the secretion of chemokines such as CCL5 and IL-8, a factor that promotes neutrophils (
49). Chemokines are produced in abundance during RSV infection in humans (
90,
115,
145), and RSV infection of BALB/c mice induces expression of CXC, CC, and C chemokines in the lung (
61,
106). Cytokine depletion, receptor blockade, or genetic deletion of chemokines or their receptors generally reduces disease severity and pathology during RSV infection. For example, antibody depletion of CCL5 (
167) or CCL11 (
101) reduces eosinophilia and disease severity in immune-augmented RSV disease, and MIP1α knockout mice have less severe disease during primary RSV infection (
61).
CCL5 (RANTES) seems of particular interest. It is produced in response to stimuli such as IFN-γ, IL-1α, IL-1β, and TNF by many cells, including fibroblasts, smooth muscle cells, and epithelial cells; in later stages of infection it is made by infiltrating cells, including γδ T cells. It selectively recruits monocytes and memory T cells and eosinophils and (at high concentrations) activates T cells. Treatment of HEp-2 cells with recombinant human CCL5 inhibits infection with RSV in vitro, an effect not seen with other chemokines. This action may result from blocking RSV fusion with host cells (
42). CCL5 increases after RSV infection of mice and correlates with the severity of disease. Anti-CCL5 antibody administration decreases airway hyperreactivity and increases IL-12 production. Moreover, CCL5 production appears to be regulated by IL-13 which is also important in RSV-induced airway hyperreactivity (
167). CCL5 may also be important in humans, since genetic studies show that polymorphisms of CCR5 affect disease severity (
74). Moreover, CCL5 levels in nasal secretions during acute RSV bronchiolitis, although not correlated with disease severity, may be predictive of the later development of recurrent wheeze (
29).
Cytokine production has also been extensively studied in RSV bronchiolitis. For example, IL-9 is a cytokine associated with Th2 responses and with asthma (
114). IL-9 mRNA and protein production is elevated in the lungs of infants with RSV bronchiolitis. Intriguingly, polymorphonuclear cells (PMN) seem to be a major source of IL-9 in this situation (
103).
RSV infection thus triggers innate immune responses that influence the developing acquired immune response (
71). NK cells are an abundant source of IFN-γ, which has potent effects on developing αβ T cells and thus on the immunopathology of RSV infection (
81). IL-12 from antigen-presenting cells has potent effects on NK cells, enhancing IFN-γ production (
78). This cascade of innate and acquired events during RSV infection is outlined in Fig.
2.
SUMMARY OF IMMUNE MECHANISMS OF RSV DISEASE
Studies of immune-augmented secondary disease are sometimes interpreted as being informative about the origins of disease in primary bronchiolitis, but the relationship between the pathogeneses of these conditions is disputed. It is important to resist overinterpreting studies of immune sensitization when attempting to explain the pathogenesis of primary bronchiolitis. However, there are common themes linking the two conditions.
The incubation period of primary RSV is about 5 days; children hospitalized with bronchiolitis have usually already been ill for 3 to 6 days. Virus replicates to higher levels and remains detectable for longer during primary infection than after prior sensitization. It is therefore possible that “acquired” T- and B-cell responses are developing by this time and contribute to disease. This situation is exacerbated by prior sensitization or in those otherwise predisposed to particular specific immune responses.
Important common and distinct factors in different forms of RSV disease are illustrated in Fig.
5. In primary infection of nonvaccinated individuals, the disease is characterized by a relatively high viral load and the delayed appearance of disease, the timing of which coincides with the development of specific acquired T-cell immunity (Fig.
5A). In previously sensitized individuals (or perhaps those in some way predisposed to enhanced disease), viral replication is more limited and viral elimination occurs more rapidly (Fig.
5B). However, the heightened immune response leads to even more disease than is typical during first infections.
Age and host genetics certainly affect the balance of immune responses during primary infection, making first RSV infections sometimes resemble secondary disease. In the neonatal period, Th1 responses are generally poor or short-lived and IL-12 production weak. One possibility is that both Th1 and Th2 responses are formed during the initial infection; however, there is specific IL-4-dependent apoptosis of Th1 cells in the neonatal environment (
94). This may explain why at this age, there is a natural tendency towards strong Th2 responses. An autocrine Th2 feedback loop driven by IL-4 might therefore tend to cause a cascade of events causing immune damage or future skewing of response to disease.
Vaccine Development and Future Therapies for RSV Disease
Despite continuous efforts to develop safe and effective vaccines, spanning 40 years, none have been successful. The basic difficulties to be overcome include the fact that natural infection gives only transient and partial immunity to reinfection of the upper respiratory tract (
64). Many established vaccines are ineffective in the first 6 months of life (
147), and it is in this period that most children suffer from RSV bronchiolitis. Vaccines would be useful in the elderly, but many established vaccines are poorly immunogenic in older people. Nonetheless, the potential benefits of an effective vaccine are undoubted. Vaccines under development include cold-passaged live viruses (
33), purified proteins (
171), and conventional inactivated and DNA vaccines (
17).
Given that disease appears to be in some part caused by immune overactivity, it is logical to test specific short-lived immune inhibitors. Steroids appear to be of little or no value, possibly because they lack specificity and reduce the severity of the immune response while potentially increasing viral replication. Anti-TNF, anti-IL-9, and anti-immunoglobulin E therapy all merit consideration, perhaps in combination with antiviral treatments. Such experimental treatments would need to be carefully justified in patients with the most severe forms of bronchiolitis. Effective prophylaxis is available in the form of a biosynthetic humanized monoclonal anti-F antibody, palivizumab. This is administered as a monthly intramuscular injection and is highly effective in preventing infection. However, its cost prohibits its use in resource-poor settings.
CONCLUSION
Severe RSV disease appears to be associated with a misdirected immune response, characterized by enhanced release of mediators and infiltration of a range of monocytes and polymorphonuclear cells. Animal models are essential to understanding disease enhancement and to the development of safe and effective vaccines, but none is ideal in all aspects (Table
1). While it is clear that primary and immune-augmented RSV diseases are not identical, these models shed light on which components of the immune response warrant further study and give rise to important general conclusions about immunopathogenesis: (i) single clinical syndromes (bronchiolitis, asthma, etc.) can result from diverse pathogenic pathways, (ii) the time and place of sampling are critical in acute transient diseases, (iii) samples from remote locations may be misleading, and (iv) the most numerous cells may not be the most influential.
The effects of formalin-inactivated RSV are remarkably similar in all species studied, including cattle, rodents, and primates. Such vaccines enhance disease by multiple pathways, leading to overactive acquired immune responses. In animal models, this disease is relatively hard to block by specific immunomodulation. By contrast, the immunopathogenesis of augmented disease in inbred mice vaccinated with single RSV proteins is highly specific and relatively easy to prevent by treatments that take out single components of the immune response (
80,
83).
It seems clear that the immunopathogenesis of RSV disease during primary infection varies considerably from one individual to another and is affected by the postnatal age. Severe RSV bronchiolitis occurs only in a small minority of children and is usually transient and self-limiting. However, studies of disease augmentation by prior sensitization are capable of reproducing many features of the severe primary disease seen in susceptible individuals and may therefore give indications of what type of immunomodulation should be attempted. A great deal has been learned in recent years, and it is to be hoped that clinical application of this knowledge will soon benefit the large numbers of children who suffer from RSV infections each year.