- Email: [email protected]
Respiratory Syncytial Virus Immunobiology and Pathogenesis
Virology 297, 1–7 (2002) doi:10.1006/viro.2002.1431
MINIREVIEW Respiratory Syncytial Virus Immunobiology and Pathogenesis Barney S. Graham, 1 John A. Rutigliano, and Teresa R. Johnson Viral Pathogenesis Laboratory, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892-3017 Received February 22, 2002; accepted February 21, 2002
though underlying asthma and chronic bronchitis can be exacerbated during infection. However, in the institutionalized elderly, disease can be severe and in some studies has been associated with as much excess mortality as influenza (Falsey et al., 1995). The pathology is less distinctive in this setting and is often complicated by chronic lung disease, heart failure, and bacterial superinfection. Therefore, it is difficult to ascribe disease pathogenesis to a direct manifestation of virus-mediated cytopathology and virus-induced inflammation, or to exacerbation of underlying conditions. RSV infection is associated with high mortality in patients with immunodeficiency. This is particularly true in the setting of severe combined immunodeficiency disease (Fishaut et al., 1980), and in patients following allogenic bone marrow (Hertz et al., 1989) or lung transplantation (Krinzman et al., 1998). Solid organ transplant recipients, patients with malignancies, or persons with other selected immune deficiencies have milder disease. The disease syndrome is a giant cell pneumonia with widespread syncytium formation in epithelial cells, presumably a direct result of virus-induced cytopathology. Finally, RSV has a legacy of vaccine-enhanced disease. Four studies using a Formalin-inactivated, alumprecipitated whole virus vaccine (FI-RSV) were performed in children in the mid-1960s. Vaccine-induced immune responses did not protect against subsequent natural infection, and disease severity was increased. Among vaccinated children in the youngest age group, 80% required hospitalization and some died (Kim et al., 1969). This tragedy has motivated extensive study of disease pathogenesis in animal models over the last 25 years and led to the current paradigms that guide vaccine development and approaches to therapy. Serological studies in the children with RSV vaccineenhanced illness showed poor neutralizing activity and fusion-inhibiting activity in the antibody response to vaccine (Murphy et al., 1986, 1988), perhaps explaining the lack of protection from natural infection. However, this
Respiratory syncytial virus (RSV) is a pneumovirus in the family Paramyxoviridae. It is a pleomorphic enveloped virus about 150 nm in diameter, transmitted by large particle aerosol and fomites, resulting in yearly epidemics of respiratory disease. RSV infects nearly 70% of infants in their first year of life and everyone by age three (Glezen and Denny, 1973). Reinfection is common. About 50% of children infected in the first year of life are reinfected in the second year of life, and everyone is reinfected every two to three years throughout life. RSV disease syndromes. There are several distinct disease syndromes associated with RSV infection. While most infants have mild disease, about 5% of those less than one year of age require hospitalization. The peak hospitalization rate is in infants between 2 and 3 months of age. Those with bronchopulmonary dysplasia and congenital heart disease are at greatest risk, but most hospitalized children have no apparent underlying condition. The virus infects both ciliated and nonciliated cells in small (75–300 m in diameter) airways, resulting in a bronchiolitis characterized clinically by wheezing. Bronchoalveolar lavage (BAL) reveals a predominance of polymorphonuclear leukocytes in respiratory secretions. However, tissue sections reveal a predominance of lymphocytes in the peribronchiolar and perivascular spaces, and small airways containing fibrin, mucus, and a mixture of sloughed epithelial cells and inflammatory cells. In addition, RSV infects the alveolar epithelium, particularly Type 1 alveolar pneumocytes. This component of RSV pathology is underappreciated despite signs of lower airway disease such as hypoxia that are commonly found even in asymptomatic patients. Most normal children and adults experience mild upper respiratory disease during reinfection with RSV, al-
1
To whom correspondence and reprint requests should be addressed at Building 40, Room 2502, 40 Convent Drive, MSC-3017, Bethesda, MD 20892-3017. Fax: (301) 480-2771. E-mail: [email protected]. 1
0042-6822/02
2
MINIREVIEW
does not explain why illness was enhanced. They were also found to have high lymphoproliferative responses to RSV antigens (Kim et al., 1976) and a relatively high frequency of eosinophilia (Chin et al., 1969). Tissue sections from the children who died demonstrate a typical distribution of inflammation around the small airways, but the composition of the infiltrate is different from primary infection with an abundance of eosinophils and neutrophils in addition to mononuclear cells (Fig. 1). There was no evidence of extensive, virus-induced cytopathology. These observations suggested a virus-specific cell-mediated immune process was responsible for the enhanced illness. The findings in animal models that relate to the syndrome of the RSV vaccine-enhanced illness will be reviewed. In particular, how the current understanding of vaccine-enhanced illness pathogenesis can inform our understanding of disease associated with primary RSV infection will be addressed. The central focus will be on how prior antigen exposure can induce aberrant adaptive immune responses, and how those responses can cause disease. The pathogenesis of RSV in the institutionalized elderly or immunocompromised patients will not be discussed in detail. Murine model of RSV. A murine model of RSV has been used to study the pathogenesis of vaccine-enhanced illness. Most work has been done in BALB/c mice that are relatively permissive to RSV replication compared to other strains. However, the murine model has some limitations that should be recognized in the interpretation of results. The infection is initiated in anesthetized mice by intranasal inoculation of a large volume of high titer virus. The inoculum is aspirated directly into lung. Virus replication occurs primarily in the alveolar parenchyma in Type 1 pneumocytes (Figs. 2A and 2B). There is evidence of bronchiolar cytopathology in some strains, and immune infiltration occurs around arterioles and bronchioles, as well as in the alveolar interstitium. Illness is mediated predominantly by the T cell response to infection (Graham et al., 1991b). Therefore, the murine model provides a good system to ask how the composition of the inflammatory infiltrate is regulated in response to RSV infection. It is not a good model for studying disease caused by virus-induced cytopathology in airway epithelium. It is also not a good system for studying the kinetics of primary RSV infection or the role of early events mediated by innate immunity, because the inoculum is so high and is delivered directly into the lung. These types of questions should be addressed in systems in which the virus is specific for the host, and a low inoculum can initiate infection in the upper respiratory tract that subsequently spreads to the lower airway. Examples include pneumonia virus of mice (PVM) or Sendai infection in mice, BRSV in calves, or RSV in humans. With these caveats in mind, using the murine model to study the composition and regulation of adaptive immune re-
sponses induced by RSV has distinct advantages because of the wealth of immunological reagents and genetically defined strains available. The Th2 hypothesis of RSV pathogenesis. In mice immunized with FI-RSV and then challenged with RSV, one sees an inflammatory infiltrate composed of monocytes, lymphocytes, and numerous eosinophils demonstrated by their large cytoplasmic granules (Fig. 2C). This is reminiscent of the pathology seen in infants with FI-RSV vaccine-enhanced disease. The pathology is dependent on CD4 ⫹ T cells. Lung extracts from mice challenged with RSV after FI-RSV immunization show a dominant IL-4 mRNA pattern, whereas the dominant response in mice immunized with live RSV given intramuscularly or intranasally is IFN-␥ (Graham et al., 1993). These findings suggested that an aberrant CD4 ⫹ Th2 response may be the basis for the RSV vaccineenhanced illness syndrome. The mechanism by which this occurs is still controversial as discussed below. When the RSV vaccine-enhanced illness was associated with CD4 ⫹ Th2 responses, it promoted an interest in the role of Type 2 cytokines in the pathogenesis of primary RSV infection. Primary RSV infection is associated with wheezing, and severe disease requiring hospitalization is associated with a higher incidence of subsequent childhood asthma (Sigurs et al., 1995). Some groups have shown RSV-specific IgE can be found in respiratory secretions of children with severe disease (Welliver et al., 1981), and others have shown an increase of eosinophilic cationic protein (Garofalo et al., 1992), findings consistent with Type 2 cytokine activity. However, others have shown children with severe disease have a dominant IFN-␥ response without direct evidence of the classical Type 2 cytokines IL-4 or IL-5 (Brandenburg et al., 2000). How is the RSV-specific Th2 response induced in FIRSV vaccine-enhanced illness? The RSV G glycoprotein is part of the envelope spike on the surface of the virus and is the putative attachment protein. G has a number of interesting structural features. It is a type II integral membrane protein anchored at the amino terminus. About 50% of the protein produced is secreted from the infected cell because of an alternative initiation codon in the transmembrane domain (Roberts et al., 1994). About 60% of the molecular weight is O-glycosylation, and there is a high serine and threonine content and a very high proline content making it more homologous to mucins than any known viral glycoprotein. The glycosylation is concentrated on either end of the molecule, and in the central portion there is a cysteine noose with a known heparan sulfate binding domain (Feldman et al., 1999) and a CX3CR binding motif (Tripp et al., 2001). RSV G has equally interesting antigenic properties. Despite careful analysis, there has never been a Gspecific CD8 ⫹ T cell response reported. The reason for this is not known. It is probably not due to a defect in
MINIREVIEW
3
FIG. 1. Haematoxylin and eosin-stained lung section from the autopsy of a child who died in the one of the original FI-RSV trials complicated by vaccine-enhanced illness. The box in (A) framing a segment of the bronchiole is shown at higher magnification in (B). One of the many eosinophils is marked by a white arrow. FIG. 2. RSV infection of BALB/c mice is restricted primarily to the alveolar epithelial cells. (A) In situ hybridization of RSV RNA in a lung section taken on day 4 after RSV challenge. (B) Transmission electron micrograph of Type 1 alveolar pneumocytes with budding filamentous RSV on day 4 of infection. The bar represents 200 nm. (C) Transmission electron micrograph of lung section from a mouse immunized with FI-RSV and subsequently challenged with RSV. Eosinophils are demonstrated by white arrows and by the presence of large cytoplasmic granules. FIG. 3. Impact of CD8 ⫹ CTL killing mechanisms on RSV pathogenesis. (A) illustrates the process of clearing RSV infection under normal circumstances. CD8 ⫹ cytotoxic T lymphocytes (CTL) recognize RSV-infected epithelial cells through T cell receptor interactions with RSV-derived peptides presented in MHC class I molecules. The majority of infected cells are eliminated via secretion of perforin and granzymes by CTL. However, FasL expression (lightning bolt) by activated CTL also participates in mediating apoptosis of RSV-infected cells. In addition, Type 1 cytokines, such as IFN-␥ and TNF-␣, are secreted by CTL in modest amounts. These combined processes are hypothesized to selectively induce apoptosis (Red X) in RSV-infected cells with very little injury of uninfected cells. (B) represents the immune response hypothesized to occur in the setting of Type 2 cytokines, particularly IL-4. Increased IL-4 leads to upregulation of FasL on activated CTL. Since most cells express a basal level of Fas, the augmented expression of FasL on CTL may result in bystander killing of uninfected cells along with killing of RSV-infected cells. In addition, the less efficient CTL produce exaggerated levels of IFN-␥ and TNF-␣. The altered CTL activity in this setting is hypothesized to increase bystander killing and to cause more severe lung pathology and enhanced illness.
4
MINIREVIEW
processing MHC class I-restricted epitopes contained within G, since the H-2K d-restricted CTL epitope from M2 inserted into G and expressed in vaccinia can induce a CTL response (Srikiatkachorn and Braciale, 1997). The lack of MHC class I-restricted response to G suggests that the memory T cell response to G is heavily biased toward MHC class II-restricted CD4 ⫹ T cells. Openshaw et al. (1992) first showed that mice immunized with recombinant vaccinia expressing RSV G develop an eosinophil response in lung and BAL following RSV challenge. Recombinant vaccinia expressing WT G, but none of the other RSV proteins, results in a significant percentage of eosinophils in BAL. They subsequently showed this response was dependent on the genetic background of the mouse (Hussell et al., 1998). Several groups then showed that G contains a dominant MHC class II-restricted epitope (aa 184–198) that induces both Th1 and Th2 CD4 ⫹ T cells (Tebbey et al., 1998; Sparer et al., 1998; Srikiatkhachorn et al., 1999). This epitope is in the region of the heparan sulfate binding domain and just downstream of the cysteine noose structure. Varga et al. then showed the CD4 ⫹ T cells responding to this epitope predominantly utilize V14 in their T cell receptors (Varga et al., 2001). These unique properties of G begged the question, were G-specific immune responses responsible for the RSV vaccine-enhanced illness? This has obvious implications for vaccine development since if true, G determinants could simply be removed from vaccine products. However, most of the antigenic variation present in RSV isolates occurs in the carboxyl-terminus of G, suggesting that G is an important target for the immune response and is under selective pressure to change. Therefore, including G in vaccine formulations would be advantageous. There are several arguments that G antigenic determinants were not the essential factor in the FI-RSV vaccine-enhanced illness. They include the following: 1. G is expressed on wild-type virus, which typically leads to a Th1-type response in primary infection. 2. The G-induced response in mice is dependent on the genetic background, is epitope dependent, and is restricted to selected T cell receptors utilizing V14. In contrast, FI-RSV immunization seemed to lead to enhanced disease in virtually all individuals, suggesting that it was not dependent on the genetic background or restricted to a single epitope or T cell receptor. 3. Pathology, illness, and immune responses similar to those elicited by FI-RSV can be elicited by purified RSV F glycoprotein, especially when formulated in alum (Graham et al., 1993; Hancock et al., 1996). However, there are other properties of G that could affect the virus-specific immune response. G is produced early in the infected cell and secreted (Hendricks et al., 1987). Therefore it is one of the first RSV proteins avail-
able in the extracellular milieu for processing through the endocytic pathway and presentation to CD4 ⫹ T cells. This may allow certain G-specific CD4 ⫹ T cell responses to occur prior to the modulating influences of IFN-␥ produced by CD8 ⫹ T cells and NK cells. In addition, the secreted G can bind to heparan sulfate moieties (Feldman et al., 1999), fractalkine receptor (Tripp et al., 2001), or lectin-like proteins (Barr et al., 2000), perhaps triggering signaling events that influence subsequent immune responses. We have asked whether the property of G secretion is important for inducing the Type 2 cytokine profile, eosinophilia, and pathology associated with the vaccine-enhanced illness. Recombinant vaccinia viruses encoding WT G (vvGwt), membrane-anchored G (vvGr), or secreted G (vvGs) were produced by Sharon Roberts in the laboratory of Dr. Gail Wertz (Roberts et al., 1994). The vaccinia constructs were inoculated by intradermal injection at the base of the tail, and RSV challenge was performed 6 weeks later. We first asked how immunization affected eosinophil responses in BAL following subsequent RSV infection. Mock immunization or vacF does not lead to postchallenge eosinophil responses. In contrast FI-RSV and vvGs induce a robust eosinophilia. Membrane-anchored G induces eosinophilia, but to a much lesser extent, and the vvGwt (a mixture of both membrane anchored and secreted G) induces an intermediate response (Johnson et al., 1998). We next asked whether the eosinophil response required IL-4, using both antibody inhibition with the anti-IL-4 monoclonal antibody, 11B.11, and IL-4 knockout mice (Johnson et al., 1999). Earlier studies had shown that FI-RSV-induced eosinophilia following subsequent challenge was IL-4 dependent (Tang et al., 1994). These studies showed that the eosinophilia induced by RSV infection following immunization with the secreted G does not require IL-4. This led us to ask whether the G-induced response was associated with IL-13, which is induced at high levels after G immunization. We used IL-4 knockout mice in combination with an IL-13 receptor antagonist. As before we showed that the G-induced eosinophilia occurred in IL-4 KO mice, whereas the FI-RSV induced eosinophilia does not occur in IL-4 KO mice. Similarly, IL-13 receptor antagonist alone did not attenuate G-induced eosinophilia, but did eliminate FI-RSV-induced eosinophilia. Interestingly, if IL-13 receptor antagonist is given to IL-4 KO mice, G is not able to induce eosinophils. This indicates that IL-4 and IL-13 are able to compensate for each other to allow G-induced eosinophilia, but both cytokines are needed for FI-RSV to induce the response (T. R. Johnson et al., submitted). These findings support the concept that G-specific epitopes were not necessarily responsible for eliciting the vaccine-enhanced illness and suggest that the pattern of immunizing parenterally with an antigen processed and presented through the endocytic pathway in
MINIREVIEW
MHC class II molecules followed by airway infection with RSV has the potential to elicit a Type 2 immune response. This sequence of parenteral immunization followed by airway challenge is known to elicit airway hypersensitivity and eosinophil responses even when using a model antigen such as ovalbumin (Willis-Karp, 1999). As mentioned above, purified F formulated in alum followed by RSV infection also results in airway eosinophilia (Graham et al., 1993). Likewise, priming mice with recombinant vaccinia expressing a secreted F glycoprotein, as opposed to the wild-type membrane-anchored F, also induces IL-4 production following RSV challenge (Bembridge et al., 1999). Therefore, any immunization strategy in RSV-naı¨ve infants based on products with obligate processing through the endocytic pathway and presentation strictly to CD4 ⫹ T lymphocytes may have the capacity to elicit the vaccine-enhanced illness. RSV has other properties that may promote Type 2 cytokine production. RSV shares a legacy of vaccineenhanced illness with measles. In both cases vaccination was with whole inactivated virus that elicited poor functional antibody responses. There is evidence in both cases that the enhanced illness was caused by an aberrant T cell response. Both viruses are in the family Paramyxoviridae and share several properties. In particular, both viruses have innate mechanisms for inhibiting the effects of IFN-␣/ (Naniche et al., 2000; Schlender et al., 2000). The NS1 and NS2 proteins of bovine RSV have been shown to antagonize the antiviral effects of IFN-␣/ (Schlender et al., 2000). Durbin et al. have also shown that the lack of IFN-␣/-mediated signaling during RSV infection promotes eosinophilia and Type 2 cytokine production (Durbin et al., 2002). It seems likely that this may be an important determinant in why RSV is associated with vaccine-enhanced illness and other disease syndromes associated with Type 2 cytokine production. However, more work is needed on how IFN-␣/ influences the differentiation of T cells, and how interference with these pathways contributes to the pathogenesis of RSV and other virus infections. Do Type 2 cytokines play a role in the pathogenesis of severe primary RSV infection? RSV infections in young infants requiring hospitalization are associated with evidence of eosinophil cationic protein in airway secretions. However, it is not clear whether disease in this setting is a manifestation of Type 2 cytokine production or not. Type 2 cytokines are not found in respiratory secretions of these infants. In contrast, large amounts of IFN-␥ are found (Brandenburg et al., 2000). It is possible that Type 2 cytokines are acting locally, but evaluation of cytokine production from peripheral blood mononuclear cells (PBMCs) of RSV-infected children (Anderson et al., 1994) and studies in animal models do not bear that out. Primary infection with RSV in mice does not induce significant levels of Type 2 cytokine production unless there is some prior stimulus such as allergic airway
5
sensitization with ovalbumin (Peebles et al., 2001a) or priming with pertussis toxin (Samore and Siber, 1996; Fischer et al., 1999). In a murine model of PVM eosinophilia is seen following primary infection, but it is not a consequence of IL-5 and the adaptive immune response. Rather, it occurs early in infection and is mediated largely by MIP-1␣ and other components of the innate immune response (Domachowske et al., 2000). This model of a pneumovirus matched to its natural host may better reflect the early events following infection than the model using human RSV in the mouse. It is possible that in children with underlying atopy or other factors predisposing to a Type 2 cytokine environment the response to primary RSV infection may be more likely to involve Th2 CD4 ⫹ T cells. This has been demonstrated in mice (Peebles et al., 2001a), but more work is needed in clinical studies to define whether this happens in humans. How do Type 2 cytokines influence disease severity during virus infection? While there are many observations of pathogens that have more severe disease in the setting of excess Type 2 cytokine production, the mechanisms of disease enhancement are poorly defined. In the airway, increased mucus production could contribute to airway obstruction, or eosinophil recruitment and degranulation could lead to direct tissue damage. Th2 CD4 ⫹ T cells are not inherently more pathogenic, and in the setting of autoimmunity they are less destructive than Th1 CD4 ⫹ T cells. In many cases it appears that the mechanisms for clearance of the pathogen are less effective, resulting in direct tissue damage from the pathogen itself. However, in the murine model of RSV and other viruses, the enhanced illness is not related to direct pathogen-induced cytopathology, but is dependent on the immune response. Therefore, we asked how IL-4 influenced the mechanisms of RSV clearance and how these immunological effectors influenced disease pathogenesis. Two early observations led us to focus on CD8 ⫹ T cells as a key mediator of this process. First, inhibiting IL-4 during FI-RSV immunization improved CD8 ⫹ T cell response following subsequent challenge and led to reduced illness (Tang and Graham, 1994). Second, CD8 ⫹ T cell-mediated RSV clearance is diminished and pathology is enhanced in IL-4 overexpressing mice (Fischer et al., 1997). These findings suggested that IL-4 may be altering functional properties of CD8 ⫹ T cells, causing them to have less efficient cytolytic function and to produce more pathology. We therefore constructed recombinant vaccinia viruses to express the RSV M2 protein that contains a defined H-2K d-restricted T cell epitope with or without coexpression of IL-4. Using these vectors, we showed that IL-4 coexpression reduced the total cytolytic capacity of the M2-specific CD8 ⫹ T cells (Aung et al., 1999). This raised the question of which cytolytic mechanisms were affected. Naı¨ve CD8 ⫹ T cells recognize processed peptides in the context of MHC class I molecules and
6
MINIREVIEW
become activated after costimulation through interaction between B7 and CD28. This leads to upregulated expression of selected molecules. When a virus-infected cell is detected by TCR interaction with the relevant epitope, perforin is released and allows the entry of granzymes and serine proteases into the target cell. This is a calcium-dependent process. In addition, FasL expression is upregulated and interaction with Fas can lead to calcium-independent signaling. Both pathways can lead to apoptosis and cell death. In general perforin-mediated lysis is more targeted and FasL-mediated lysis is less specific and more likely to result in antigen-independent bystander killing. We found that coexpression of IL-4 with RSV M2 in recombinant vaccinia increases FasL expression on the M2-specific CD8 ⫹ T cells resulting in diminished perforin-mediated cytolytic activity and enhanced FasL-mediated target cell killing (Aung and Graham, 2000). Experiments in perforin KO mice confirmed that perforin was not necessary for RSV clearance, but when other cytolytic mechanisms are used, clearance and disease expression are delayed, and illness is more prolonged. In addition, excess production of IFN-␥ and TNF-␣ in lung temporally correlated with the illness (Aung et al., 2001). In contrast, when FasL-defective (gld) mice are infected with RSV, viral clearance occurs with normal kinetics with minimal illness and normal amounts of IFN-␥ and TNF-␣ (J. A. Rutigliano, unpublished observations). These observations suggest a new working hypothesis for the mechanism of severe illness in RSV. We propose that the critical pathogenic event that causes illness in primary RSV infection is an aberrant CD8 ⫹ T cell response. Specifically, inefficient cytolytic mechanisms will lead to delayed and prolonged virus clearance, and bystander injury may be the basis of enhanced pathology in settings where too much reliance is placed on FasL-mediated mechanisms of lysis and cytotoxic cytokines (Fig. 3). An important corollary of this hypothesis is that immunization approaches that induce subsets of CD8 ⫹ CTL responses with efficient, targeted killing mechanisms should result in more favorable outcomes. These concepts will be the subject of future studies in murine models and clinical trials. REFERENCES Anderson, L. J., Tsou, C., Potter, C., Keyserling, H. L., Smith, T. F., Ananaba, G., and Bangham, C. R. (1994). Cytokine response to respiratory syncytial virus stimulation of human peripheral blood mononuclear cells. J. Infect. Dis. 170, 1201–1208. Aung, S., and Graham, B. S. (2000). Differential regulation of perforinand FasL-mediated cytotoxicity by IL-4. J. Immunol. 164, 3487–3493. Aung, S., Rutigliano, J. A., and Graham, B. S. (2001). Alternative mechanisms of respiratory syncytial virus clearance in perforin knockout mice lead to enhanced disease. J. Virol. 75, 9918–9924. Aung, S., Tang, Y. W., and Graham, B. S. (1999). IL-4 inhibits induction of cytotoxic T lymphocyte activity in mice infected with recombinant
vaccinia virus expressing respiratory syncytial virus M2 protein. J. Virol. 73, 8944–8949. Barr, F. E., Pedigo, H., Johnson, T. R., and Shepherd, V. L. (2000). Surfactant protein-A enhances uptake of respiratory syncytial virus by monocytes and U937 macrophages. Am. J. Respir. Cell Mol. Biol. 23, 586–592. Bembridge, G. P., Lopez, J. A., Bustos, R., Melero, J. A., Cook, R., Mason, H., and Taylor, G. (1999). Priming with a secreted form of the fusion protein of respiratory syncytial virus (RSV) promotes interleukin-4 (IL-4) and IL-5 production but not pulmonary eosinophilia following RSV challenge. J. Virol. 73, 10086–10094. Brandenburg, A. H., Kleinjan, A., van Het Land, B., Moll, H. A., Timmerman, H. H., de Swart, R. L., Neijens, H. J., Fokkens, W., and Osterhaus, A. D. (2000). Type 1-like immune response is found in children with respiratory syncytial virus infection regardless of clinical severity. J. Med. Virol. 62, 267–277. Chin, J., Magoffin, R. L., Shearer, L. A., Schieble, J. H., and Lennette, E. H. (1969). Field evaluation of a respiratory syncytial virus vaccine and a trivalent parainfluenza virus vaccine in a pediatric population. Am. J. Epidemiol. 89, 449–463. Domachowske, J. B., Bonville, C. A., Dyer, K. D., Easton, A. J., and Rosenberg, H. F. (2000). Pulmonary eosinophilia and production of MIP-1alpha are prominent responses to infection with pneumonia virus of mice. Cell Immunol. 200, 98–104. Durbin, J., Johnson, T. R., Durbin, R. K., Mertz, S., Peebles, R. S., Jr., and Graham, B. S. (2002). The role of interferon in RSV pathogenesis. J. Immunol. 168, 2944–2951. Falsey, A. R., Cunningham, C. K., Barker, W. H., Kouides, R. W., Yuen, J. B., Menegus, M., Weiner, L. B., Bonville, C. A., and Betts, R. F. (1995). Respiratory syncytial virus and influenza A infections in the hospitalized elderly. J. Infect. Dis. 172, 389–394. Feldman, S. A., Hendry, R. M., and Beeler, J. A. (1999). Identification of a linear heparin binding domain for human respiratory syncytial virus attachment glycoprotein G. J. Virol. 73, 6610–6617. Fishaut, M., Tubergen, D., and McIntosh, K. (1980). Cellular response to respiratory viruses with particular reference to children with disorders of cell-mediated immunity. J. Pediatr. 96, 179–186. Fischer, J. E., Johnson, J. E., Kuli-Zade, R., Johnson, T. R., Aung, S., Parker, R. A., and Graham, B. S. (1997). Overexpression of IL-4 delays virus clearance in mice infected with RSV. J. Virol. 71, 8672–8677. Fischer, J. E., Johnson, T. R., Peebles, R. S., and Graham, B. S. (1999). Vaccination with pertussis toxin alters the antibody response to simultaneous respiratory syncytial virus challenge. J. Infect. Dis. 180, 714–771. Garofalo, R., Kimpen, J. L. L., Welliver, R. C., and Ogra, P. L. (1992). Eosinophil degranulation in the respiratory tract during naturally acquired respiratory syncytial virus infection. J. Pediatr. 120, 28–32. Glezen, P., and Denny, F. W. (1973). Epidemiology of acute lower respiratory disease in children. N. Engl. J. Med. 288, 498–505. Graham, B. S., Bunton, L. A., Wright, P. F., and Karzon, D. T. (1991b). Role of T lymphocyte subsets in the pathogenesis of primary infection and rechallenge with respiratory syncytial virus in mice. J. Clin. Invest. 88, 1026–1033. Graham, B. S., Henderson, G. S., Tang, Y. W., Lu, X., Neuzil, K. M., and Colley, D. G. (1993). Priming immunization determines T helper cytokine mRNA expression patterns in lungs of mice challenged with respiratory syncytial virus. J. Immunol. 151, 2032–2040. Hancock, G. E., Speelman, D. J., Heers, K., Bortell, E., Smith, J., and Cosco, C. (1996). Generation of atypical pulmonary inflammatory responses in BALB/c mice after immunization with the native attachment (G) glycoprotein of respiratory syncytial virus. J. Virol. 70, 7783– 7791. Hendricks, D. A., Baradaran, K., McIntosh, K., and Patterson, J. L. (1987). Appearance of a soluble form of the G protein of respiratory syncytial virus in fluids of infected cells. J. Gen. Virol. 68, 1705–1714. Hertz, M. I., Englund, J. A., Snover, D., Bitterman, P. B., and McGlave, P. B. (1989). Respiratory syncytial virus-induced acute lung injury in
MINIREVIEW adult patients with bone marrow transplants: A clinical approach and review of the literature. Medicine 68, 269–281. Hussell, T., Baldwin, C. J., O’Garra, A., and Openshaw, P. J. M. (1997). CD8⫹ T cells control Th2-driven pathology during pulmonary respiratory syncytial virus infection. Eur. J. Immunol. 27, 3341–3349. Hussell, T., Georgiou, A., Sparer, T. E., Matthews, S., Pala, P., and Openshaw, P. J. (1998). Host genetic determinants of vaccine-induced eosinophilia during respiratory syncytial virus infection. J. Immunol. 161, 6215–6222. Hussell, T., and Openshaw, P. J. (1998). Intracellular IFN-gamma expression in natural killer cells precedes lung CD8⫹ T cell recruitment during respiratory syncytial virus infection. J. Gen. Virol. 79, 2593– 2601. Johnson, T. R., and Graham, B. S. (1999). Secreted respiratory syncytial virus (RSV) G protein induces IL-5, IL-13, and eosinophilia by an IL-4-independent mechanism. J. Virol. 73, 8485–8495. Johnson, T. R., Johnson, J. E., Roberts, S. R., Wertz, G. W., Parker, R. A., and Graham, B. S. (1998). Priming with secreted glycoprotein G of respiratory syncytial virus (RSV) augments interleukin-5 production and tissue eosinophilia after RSV challenge. J. Virol. 72, 2871–2880. Johnson, T. R., Parker, R. A., Johnson, J. E., and Graham, B. S. IL-1 is sufficient for respiratory syncytial virus (RSV) G glycoprotein-induced eosinophilia following RSV challenge. J. Immunol., submitted. Kim, H. W., Canchola, J. G., Brandt, C. D., Pyles, G., Chanock, R. M., Jensen, K., and Parrott, R. H. (1969). Respiratory syncytial virus disease in infants despite prior administration of antigenic inactivated vaccine. Am. J. Epidemiol. 89, 422–434. Kim, H. W., Leikin, S. L., Arrobio, J., Brandt, C. D., Chanock, R. M., and Parrott, R. H. (1976). Cell-mediated immunity to respiratory syncytial virus induced by inactivated vaccine or by infection. Pediatr. Res. 10, 75–78. Krinzman, S., Basgoz, N., Kradin, R., Shepard, J. A., Flieder, D. B., Wright, C. D., Wain, J. C., and Ginns, L. C. (1998). Respiratory syncytial virus-associated infections in adult recipients of solid organ transplants. J. Heart Lung Transplant. 17, 202–210. Lambert, D. M. (1988). Role of oligosaccharides in the structure and function of respiratory syncytial virus glycoproteins. Virology 164, 458–466. Murphy, B. R., Prince, G. A., Walsh, E. E., Kim, H. W., Parrott, R. H., Hemming, V. G., Rodriguez, W. J., and Chanock, R. M. (1986). Dissociation between serum neutralizing and glycoprotein antibody responses of infants and children who received inactivated respiratory syncytial virus vaccine. J. Clin. Microbiol. 124, 197–202. Murphy, B. R., and Walsh, E. E. (1988). Formalin-inactivated respiratory syncytial virus vaccine induces antibodies to the fusion glycoprotein that are deficient in fusion-inhibiting activity. J. Clin. Microbiol. 26, 1595–1597. Naniche, D., Yeh, A., Eto, D., Manchester, M., Friedman, R. M., and Oldstone, M. B. (2000). Evasion of host defenses by measles virus: Wild-type measles virus infection interferes with induction of alpha/ beta interferon production. J. Virol. 74, 7478–7484. Openshaw, P. J. M., Clarke, S. L., and Record, F. M. (1992). Pulmonary eosinophilic response to respiratory syncytial virus infection in mice sensitized to the major surface glycoprotein G. Inter. Immunol. 4, 493–500. Peebles, R. S., Sheller, J. R., Collins, R. D., Jarzecka, A. K., Mitchell, D. B., Parker, R. A., and Graham, B. S. (2001a). Respiratory syncytial virus
7
infection decreases allergen-induced type 2 cytokine production while increasing airway hyperresponsiveness in mice. J. Med. Virol. 63, 178–188. Peebles, R. S., Hashimoto, K., Collins, R. D., Jarzecka, K., Furlong, J., Mitchell, D. B., Sheller, J. R., and Graham, B. S. (2001b). Immune interaction between respiratory syncytial virus infection and allergen sensitization critically depends on timing of challenges. J. Infect. Dis. 184, 1374–1379. Roberts, S. R., Lichtenstein, D., Ball, L. A., and Wertz, G. W. (1994). The membrane-associated and secreted forms of the respiratory syncytial virus attachment glycoprotein G are synthesized from alternative initiation codons. J. Virol. 68, 4538–4546. Samore, M. H., and Siber, G. R. (1996). Pertussis toxin enhanced IgG1 and IgE responses to primary tetanus immunization are mediated by interleukin-4 and persist during secondary responses to tetanus alone. Vaccine 14, 290–297. Schlender, J., Bossert, B., Buchholz, U., and Conzelmann, K. K. (2000). Bovine respiratory syncytial virus nonstructural proteins NS1 and NS2 cooperatively antagonize alpha/beta interferon-induced antiviral response. J. Virol. 74, 8234–8242. Sigurs, N., Bjarnason, R., Sigurbergsson, F., Kjellman, B., and Bjorksten, B. (1995). Asthma and immunoglobulin E antibodies after respiratory syncytial virus bronchiolitis: A prospective cohort study with matched controls. Pediatrics 95, 500–505. Sparer, T. E., Matthews, S., Hussell, T., Rae, A. J., Garcia-Barreno, B., Melero, J. A., and Openshaw, P. J. M. (1998). Eliminating a region of respiratory syncytial virus attachment protein allows induction of protective immunity without vaccine-enhanced lung eosinophilia. J. Exp. Med. 187, 1921–1926. Srikiatkachorn, A., and Braciale, T. J. (1997). Virus-specific CD8⫹ T lymphocytes downregulate T helper cell type 2 cytokine secretion and pulmonary eosinophilia during experimental murine respiratory syncytial virus infection. J. Exp. Med. 186, 421–432. Srikiatkhachorn, A., Chang, W., and Braciale, T. J. (1999). Induction of Th-1 and Th-2 responses by respiratory syncytial virus attachment glycoprotein is epitope and major histocompatibility complex independent. J. Virol. 73, 6590–6597. Tang, Y. W., and Graham, B. S. (1994). Anti-IL-4 treatment at immunization modulates cytokine expression, reduces illness, and increases cytotoxic T lymphocyte activity in mice challenged with respiratory syncytial virus. J. Clin. Invest. 94, 1953–1958. Tebbey, P. W., Hagen, M., and Hancock, G. E. (1998). Atypical pulmonary eosinophilia is mediated by a specific amino acid sequence of the attachment (G) protein of respiratory syncytial virus. J. Exp. Med. 188, 1967–1972. Tripp, R. A., Jones, L. P., Haynes, L. M., Zheng, H., Murphy, P. M., and Anderson, L. J. (2001). CX3C chemokine mimicry by respiratory syncytial virus G glycoprotein. Nat. Immunol. 2, 732–738. Varga, S. M., Wang, X., Welsh, R. M., and Braciale, T. J. (2001). Immunopathology in RSV infection is mediated by a discrete oligoclonal subset of antigen-specific CD4(⫹) T cells. Immunity 15, 637–646. Welliver, R. C., Wong, D. T., Sun, M., Middleton, E., Jr., Vaughan, R. S., and Ogra, P. L. (1981). The development of respiratory syncytial virus-specific IgE and the release of histamine in nasopharyngeal secretions after infection. N. Engl. J. Med. 305, 841–884. Wills-Karp, M. (1999). Immunologic basis of antigen-induced airway hyperresponsiveness. Annu. Rev. Immunol. 17, 255–281.
MINIREVIEW Respiratory Syncytial Virus Immunobiology and Pathogenesis Barney S. Graham, 1 John A. Rutigliano, and Teresa R. Johnson Viral Pathogenesis Laboratory, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892-3017 Received February 22, 2002; accepted February 21, 2002
though underlying asthma and chronic bronchitis can be exacerbated during infection. However, in the institutionalized elderly, disease can be severe and in some studies has been associated with as much excess mortality as influenza (Falsey et al., 1995). The pathology is less distinctive in this setting and is often complicated by chronic lung disease, heart failure, and bacterial superinfection. Therefore, it is difficult to ascribe disease pathogenesis to a direct manifestation of virus-mediated cytopathology and virus-induced inflammation, or to exacerbation of underlying conditions. RSV infection is associated with high mortality in patients with immunodeficiency. This is particularly true in the setting of severe combined immunodeficiency disease (Fishaut et al., 1980), and in patients following allogenic bone marrow (Hertz et al., 1989) or lung transplantation (Krinzman et al., 1998). Solid organ transplant recipients, patients with malignancies, or persons with other selected immune deficiencies have milder disease. The disease syndrome is a giant cell pneumonia with widespread syncytium formation in epithelial cells, presumably a direct result of virus-induced cytopathology. Finally, RSV has a legacy of vaccine-enhanced disease. Four studies using a Formalin-inactivated, alumprecipitated whole virus vaccine (FI-RSV) were performed in children in the mid-1960s. Vaccine-induced immune responses did not protect against subsequent natural infection, and disease severity was increased. Among vaccinated children in the youngest age group, 80% required hospitalization and some died (Kim et al., 1969). This tragedy has motivated extensive study of disease pathogenesis in animal models over the last 25 years and led to the current paradigms that guide vaccine development and approaches to therapy. Serological studies in the children with RSV vaccineenhanced illness showed poor neutralizing activity and fusion-inhibiting activity in the antibody response to vaccine (Murphy et al., 1986, 1988), perhaps explaining the lack of protection from natural infection. However, this
Respiratory syncytial virus (RSV) is a pneumovirus in the family Paramyxoviridae. It is a pleomorphic enveloped virus about 150 nm in diameter, transmitted by large particle aerosol and fomites, resulting in yearly epidemics of respiratory disease. RSV infects nearly 70% of infants in their first year of life and everyone by age three (Glezen and Denny, 1973). Reinfection is common. About 50% of children infected in the first year of life are reinfected in the second year of life, and everyone is reinfected every two to three years throughout life. RSV disease syndromes. There are several distinct disease syndromes associated with RSV infection. While most infants have mild disease, about 5% of those less than one year of age require hospitalization. The peak hospitalization rate is in infants between 2 and 3 months of age. Those with bronchopulmonary dysplasia and congenital heart disease are at greatest risk, but most hospitalized children have no apparent underlying condition. The virus infects both ciliated and nonciliated cells in small (75–300 m in diameter) airways, resulting in a bronchiolitis characterized clinically by wheezing. Bronchoalveolar lavage (BAL) reveals a predominance of polymorphonuclear leukocytes in respiratory secretions. However, tissue sections reveal a predominance of lymphocytes in the peribronchiolar and perivascular spaces, and small airways containing fibrin, mucus, and a mixture of sloughed epithelial cells and inflammatory cells. In addition, RSV infects the alveolar epithelium, particularly Type 1 alveolar pneumocytes. This component of RSV pathology is underappreciated despite signs of lower airway disease such as hypoxia that are commonly found even in asymptomatic patients. Most normal children and adults experience mild upper respiratory disease during reinfection with RSV, al-
1
To whom correspondence and reprint requests should be addressed at Building 40, Room 2502, 40 Convent Drive, MSC-3017, Bethesda, MD 20892-3017. Fax: (301) 480-2771. E-mail: [email protected]. 1
0042-6822/02
2
MINIREVIEW
does not explain why illness was enhanced. They were also found to have high lymphoproliferative responses to RSV antigens (Kim et al., 1976) and a relatively high frequency of eosinophilia (Chin et al., 1969). Tissue sections from the children who died demonstrate a typical distribution of inflammation around the small airways, but the composition of the infiltrate is different from primary infection with an abundance of eosinophils and neutrophils in addition to mononuclear cells (Fig. 1). There was no evidence of extensive, virus-induced cytopathology. These observations suggested a virus-specific cell-mediated immune process was responsible for the enhanced illness. The findings in animal models that relate to the syndrome of the RSV vaccine-enhanced illness will be reviewed. In particular, how the current understanding of vaccine-enhanced illness pathogenesis can inform our understanding of disease associated with primary RSV infection will be addressed. The central focus will be on how prior antigen exposure can induce aberrant adaptive immune responses, and how those responses can cause disease. The pathogenesis of RSV in the institutionalized elderly or immunocompromised patients will not be discussed in detail. Murine model of RSV. A murine model of RSV has been used to study the pathogenesis of vaccine-enhanced illness. Most work has been done in BALB/c mice that are relatively permissive to RSV replication compared to other strains. However, the murine model has some limitations that should be recognized in the interpretation of results. The infection is initiated in anesthetized mice by intranasal inoculation of a large volume of high titer virus. The inoculum is aspirated directly into lung. Virus replication occurs primarily in the alveolar parenchyma in Type 1 pneumocytes (Figs. 2A and 2B). There is evidence of bronchiolar cytopathology in some strains, and immune infiltration occurs around arterioles and bronchioles, as well as in the alveolar interstitium. Illness is mediated predominantly by the T cell response to infection (Graham et al., 1991b). Therefore, the murine model provides a good system to ask how the composition of the inflammatory infiltrate is regulated in response to RSV infection. It is not a good model for studying disease caused by virus-induced cytopathology in airway epithelium. It is also not a good system for studying the kinetics of primary RSV infection or the role of early events mediated by innate immunity, because the inoculum is so high and is delivered directly into the lung. These types of questions should be addressed in systems in which the virus is specific for the host, and a low inoculum can initiate infection in the upper respiratory tract that subsequently spreads to the lower airway. Examples include pneumonia virus of mice (PVM) or Sendai infection in mice, BRSV in calves, or RSV in humans. With these caveats in mind, using the murine model to study the composition and regulation of adaptive immune re-
sponses induced by RSV has distinct advantages because of the wealth of immunological reagents and genetically defined strains available. The Th2 hypothesis of RSV pathogenesis. In mice immunized with FI-RSV and then challenged with RSV, one sees an inflammatory infiltrate composed of monocytes, lymphocytes, and numerous eosinophils demonstrated by their large cytoplasmic granules (Fig. 2C). This is reminiscent of the pathology seen in infants with FI-RSV vaccine-enhanced disease. The pathology is dependent on CD4 ⫹ T cells. Lung extracts from mice challenged with RSV after FI-RSV immunization show a dominant IL-4 mRNA pattern, whereas the dominant response in mice immunized with live RSV given intramuscularly or intranasally is IFN-␥ (Graham et al., 1993). These findings suggested that an aberrant CD4 ⫹ Th2 response may be the basis for the RSV vaccineenhanced illness syndrome. The mechanism by which this occurs is still controversial as discussed below. When the RSV vaccine-enhanced illness was associated with CD4 ⫹ Th2 responses, it promoted an interest in the role of Type 2 cytokines in the pathogenesis of primary RSV infection. Primary RSV infection is associated with wheezing, and severe disease requiring hospitalization is associated with a higher incidence of subsequent childhood asthma (Sigurs et al., 1995). Some groups have shown RSV-specific IgE can be found in respiratory secretions of children with severe disease (Welliver et al., 1981), and others have shown an increase of eosinophilic cationic protein (Garofalo et al., 1992), findings consistent with Type 2 cytokine activity. However, others have shown children with severe disease have a dominant IFN-␥ response without direct evidence of the classical Type 2 cytokines IL-4 or IL-5 (Brandenburg et al., 2000). How is the RSV-specific Th2 response induced in FIRSV vaccine-enhanced illness? The RSV G glycoprotein is part of the envelope spike on the surface of the virus and is the putative attachment protein. G has a number of interesting structural features. It is a type II integral membrane protein anchored at the amino terminus. About 50% of the protein produced is secreted from the infected cell because of an alternative initiation codon in the transmembrane domain (Roberts et al., 1994). About 60% of the molecular weight is O-glycosylation, and there is a high serine and threonine content and a very high proline content making it more homologous to mucins than any known viral glycoprotein. The glycosylation is concentrated on either end of the molecule, and in the central portion there is a cysteine noose with a known heparan sulfate binding domain (Feldman et al., 1999) and a CX3CR binding motif (Tripp et al., 2001). RSV G has equally interesting antigenic properties. Despite careful analysis, there has never been a Gspecific CD8 ⫹ T cell response reported. The reason for this is not known. It is probably not due to a defect in
MINIREVIEW
3
FIG. 1. Haematoxylin and eosin-stained lung section from the autopsy of a child who died in the one of the original FI-RSV trials complicated by vaccine-enhanced illness. The box in (A) framing a segment of the bronchiole is shown at higher magnification in (B). One of the many eosinophils is marked by a white arrow. FIG. 2. RSV infection of BALB/c mice is restricted primarily to the alveolar epithelial cells. (A) In situ hybridization of RSV RNA in a lung section taken on day 4 after RSV challenge. (B) Transmission electron micrograph of Type 1 alveolar pneumocytes with budding filamentous RSV on day 4 of infection. The bar represents 200 nm. (C) Transmission electron micrograph of lung section from a mouse immunized with FI-RSV and subsequently challenged with RSV. Eosinophils are demonstrated by white arrows and by the presence of large cytoplasmic granules. FIG. 3. Impact of CD8 ⫹ CTL killing mechanisms on RSV pathogenesis. (A) illustrates the process of clearing RSV infection under normal circumstances. CD8 ⫹ cytotoxic T lymphocytes (CTL) recognize RSV-infected epithelial cells through T cell receptor interactions with RSV-derived peptides presented in MHC class I molecules. The majority of infected cells are eliminated via secretion of perforin and granzymes by CTL. However, FasL expression (lightning bolt) by activated CTL also participates in mediating apoptosis of RSV-infected cells. In addition, Type 1 cytokines, such as IFN-␥ and TNF-␣, are secreted by CTL in modest amounts. These combined processes are hypothesized to selectively induce apoptosis (Red X) in RSV-infected cells with very little injury of uninfected cells. (B) represents the immune response hypothesized to occur in the setting of Type 2 cytokines, particularly IL-4. Increased IL-4 leads to upregulation of FasL on activated CTL. Since most cells express a basal level of Fas, the augmented expression of FasL on CTL may result in bystander killing of uninfected cells along with killing of RSV-infected cells. In addition, the less efficient CTL produce exaggerated levels of IFN-␥ and TNF-␣. The altered CTL activity in this setting is hypothesized to increase bystander killing and to cause more severe lung pathology and enhanced illness.
4
MINIREVIEW
processing MHC class I-restricted epitopes contained within G, since the H-2K d-restricted CTL epitope from M2 inserted into G and expressed in vaccinia can induce a CTL response (Srikiatkachorn and Braciale, 1997). The lack of MHC class I-restricted response to G suggests that the memory T cell response to G is heavily biased toward MHC class II-restricted CD4 ⫹ T cells. Openshaw et al. (1992) first showed that mice immunized with recombinant vaccinia expressing RSV G develop an eosinophil response in lung and BAL following RSV challenge. Recombinant vaccinia expressing WT G, but none of the other RSV proteins, results in a significant percentage of eosinophils in BAL. They subsequently showed this response was dependent on the genetic background of the mouse (Hussell et al., 1998). Several groups then showed that G contains a dominant MHC class II-restricted epitope (aa 184–198) that induces both Th1 and Th2 CD4 ⫹ T cells (Tebbey et al., 1998; Sparer et al., 1998; Srikiatkhachorn et al., 1999). This epitope is in the region of the heparan sulfate binding domain and just downstream of the cysteine noose structure. Varga et al. then showed the CD4 ⫹ T cells responding to this epitope predominantly utilize V14 in their T cell receptors (Varga et al., 2001). These unique properties of G begged the question, were G-specific immune responses responsible for the RSV vaccine-enhanced illness? This has obvious implications for vaccine development since if true, G determinants could simply be removed from vaccine products. However, most of the antigenic variation present in RSV isolates occurs in the carboxyl-terminus of G, suggesting that G is an important target for the immune response and is under selective pressure to change. Therefore, including G in vaccine formulations would be advantageous. There are several arguments that G antigenic determinants were not the essential factor in the FI-RSV vaccine-enhanced illness. They include the following: 1. G is expressed on wild-type virus, which typically leads to a Th1-type response in primary infection. 2. The G-induced response in mice is dependent on the genetic background, is epitope dependent, and is restricted to selected T cell receptors utilizing V14. In contrast, FI-RSV immunization seemed to lead to enhanced disease in virtually all individuals, suggesting that it was not dependent on the genetic background or restricted to a single epitope or T cell receptor. 3. Pathology, illness, and immune responses similar to those elicited by FI-RSV can be elicited by purified RSV F glycoprotein, especially when formulated in alum (Graham et al., 1993; Hancock et al., 1996). However, there are other properties of G that could affect the virus-specific immune response. G is produced early in the infected cell and secreted (Hendricks et al., 1987). Therefore it is one of the first RSV proteins avail-
able in the extracellular milieu for processing through the endocytic pathway and presentation to CD4 ⫹ T cells. This may allow certain G-specific CD4 ⫹ T cell responses to occur prior to the modulating influences of IFN-␥ produced by CD8 ⫹ T cells and NK cells. In addition, the secreted G can bind to heparan sulfate moieties (Feldman et al., 1999), fractalkine receptor (Tripp et al., 2001), or lectin-like proteins (Barr et al., 2000), perhaps triggering signaling events that influence subsequent immune responses. We have asked whether the property of G secretion is important for inducing the Type 2 cytokine profile, eosinophilia, and pathology associated with the vaccine-enhanced illness. Recombinant vaccinia viruses encoding WT G (vvGwt), membrane-anchored G (vvGr), or secreted G (vvGs) were produced by Sharon Roberts in the laboratory of Dr. Gail Wertz (Roberts et al., 1994). The vaccinia constructs were inoculated by intradermal injection at the base of the tail, and RSV challenge was performed 6 weeks later. We first asked how immunization affected eosinophil responses in BAL following subsequent RSV infection. Mock immunization or vacF does not lead to postchallenge eosinophil responses. In contrast FI-RSV and vvGs induce a robust eosinophilia. Membrane-anchored G induces eosinophilia, but to a much lesser extent, and the vvGwt (a mixture of both membrane anchored and secreted G) induces an intermediate response (Johnson et al., 1998). We next asked whether the eosinophil response required IL-4, using both antibody inhibition with the anti-IL-4 monoclonal antibody, 11B.11, and IL-4 knockout mice (Johnson et al., 1999). Earlier studies had shown that FI-RSV-induced eosinophilia following subsequent challenge was IL-4 dependent (Tang et al., 1994). These studies showed that the eosinophilia induced by RSV infection following immunization with the secreted G does not require IL-4. This led us to ask whether the G-induced response was associated with IL-13, which is induced at high levels after G immunization. We used IL-4 knockout mice in combination with an IL-13 receptor antagonist. As before we showed that the G-induced eosinophilia occurred in IL-4 KO mice, whereas the FI-RSV induced eosinophilia does not occur in IL-4 KO mice. Similarly, IL-13 receptor antagonist alone did not attenuate G-induced eosinophilia, but did eliminate FI-RSV-induced eosinophilia. Interestingly, if IL-13 receptor antagonist is given to IL-4 KO mice, G is not able to induce eosinophils. This indicates that IL-4 and IL-13 are able to compensate for each other to allow G-induced eosinophilia, but both cytokines are needed for FI-RSV to induce the response (T. R. Johnson et al., submitted). These findings support the concept that G-specific epitopes were not necessarily responsible for eliciting the vaccine-enhanced illness and suggest that the pattern of immunizing parenterally with an antigen processed and presented through the endocytic pathway in
MINIREVIEW
MHC class II molecules followed by airway infection with RSV has the potential to elicit a Type 2 immune response. This sequence of parenteral immunization followed by airway challenge is known to elicit airway hypersensitivity and eosinophil responses even when using a model antigen such as ovalbumin (Willis-Karp, 1999). As mentioned above, purified F formulated in alum followed by RSV infection also results in airway eosinophilia (Graham et al., 1993). Likewise, priming mice with recombinant vaccinia expressing a secreted F glycoprotein, as opposed to the wild-type membrane-anchored F, also induces IL-4 production following RSV challenge (Bembridge et al., 1999). Therefore, any immunization strategy in RSV-naı¨ve infants based on products with obligate processing through the endocytic pathway and presentation strictly to CD4 ⫹ T lymphocytes may have the capacity to elicit the vaccine-enhanced illness. RSV has other properties that may promote Type 2 cytokine production. RSV shares a legacy of vaccineenhanced illness with measles. In both cases vaccination was with whole inactivated virus that elicited poor functional antibody responses. There is evidence in both cases that the enhanced illness was caused by an aberrant T cell response. Both viruses are in the family Paramyxoviridae and share several properties. In particular, both viruses have innate mechanisms for inhibiting the effects of IFN-␣/ (Naniche et al., 2000; Schlender et al., 2000). The NS1 and NS2 proteins of bovine RSV have been shown to antagonize the antiviral effects of IFN-␣/ (Schlender et al., 2000). Durbin et al. have also shown that the lack of IFN-␣/-mediated signaling during RSV infection promotes eosinophilia and Type 2 cytokine production (Durbin et al., 2002). It seems likely that this may be an important determinant in why RSV is associated with vaccine-enhanced illness and other disease syndromes associated with Type 2 cytokine production. However, more work is needed on how IFN-␣/ influences the differentiation of T cells, and how interference with these pathways contributes to the pathogenesis of RSV and other virus infections. Do Type 2 cytokines play a role in the pathogenesis of severe primary RSV infection? RSV infections in young infants requiring hospitalization are associated with evidence of eosinophil cationic protein in airway secretions. However, it is not clear whether disease in this setting is a manifestation of Type 2 cytokine production or not. Type 2 cytokines are not found in respiratory secretions of these infants. In contrast, large amounts of IFN-␥ are found (Brandenburg et al., 2000). It is possible that Type 2 cytokines are acting locally, but evaluation of cytokine production from peripheral blood mononuclear cells (PBMCs) of RSV-infected children (Anderson et al., 1994) and studies in animal models do not bear that out. Primary infection with RSV in mice does not induce significant levels of Type 2 cytokine production unless there is some prior stimulus such as allergic airway
5
sensitization with ovalbumin (Peebles et al., 2001a) or priming with pertussis toxin (Samore and Siber, 1996; Fischer et al., 1999). In a murine model of PVM eosinophilia is seen following primary infection, but it is not a consequence of IL-5 and the adaptive immune response. Rather, it occurs early in infection and is mediated largely by MIP-1␣ and other components of the innate immune response (Domachowske et al., 2000). This model of a pneumovirus matched to its natural host may better reflect the early events following infection than the model using human RSV in the mouse. It is possible that in children with underlying atopy or other factors predisposing to a Type 2 cytokine environment the response to primary RSV infection may be more likely to involve Th2 CD4 ⫹ T cells. This has been demonstrated in mice (Peebles et al., 2001a), but more work is needed in clinical studies to define whether this happens in humans. How do Type 2 cytokines influence disease severity during virus infection? While there are many observations of pathogens that have more severe disease in the setting of excess Type 2 cytokine production, the mechanisms of disease enhancement are poorly defined. In the airway, increased mucus production could contribute to airway obstruction, or eosinophil recruitment and degranulation could lead to direct tissue damage. Th2 CD4 ⫹ T cells are not inherently more pathogenic, and in the setting of autoimmunity they are less destructive than Th1 CD4 ⫹ T cells. In many cases it appears that the mechanisms for clearance of the pathogen are less effective, resulting in direct tissue damage from the pathogen itself. However, in the murine model of RSV and other viruses, the enhanced illness is not related to direct pathogen-induced cytopathology, but is dependent on the immune response. Therefore, we asked how IL-4 influenced the mechanisms of RSV clearance and how these immunological effectors influenced disease pathogenesis. Two early observations led us to focus on CD8 ⫹ T cells as a key mediator of this process. First, inhibiting IL-4 during FI-RSV immunization improved CD8 ⫹ T cell response following subsequent challenge and led to reduced illness (Tang and Graham, 1994). Second, CD8 ⫹ T cell-mediated RSV clearance is diminished and pathology is enhanced in IL-4 overexpressing mice (Fischer et al., 1997). These findings suggested that IL-4 may be altering functional properties of CD8 ⫹ T cells, causing them to have less efficient cytolytic function and to produce more pathology. We therefore constructed recombinant vaccinia viruses to express the RSV M2 protein that contains a defined H-2K d-restricted T cell epitope with or without coexpression of IL-4. Using these vectors, we showed that IL-4 coexpression reduced the total cytolytic capacity of the M2-specific CD8 ⫹ T cells (Aung et al., 1999). This raised the question of which cytolytic mechanisms were affected. Naı¨ve CD8 ⫹ T cells recognize processed peptides in the context of MHC class I molecules and
6
MINIREVIEW
become activated after costimulation through interaction between B7 and CD28. This leads to upregulated expression of selected molecules. When a virus-infected cell is detected by TCR interaction with the relevant epitope, perforin is released and allows the entry of granzymes and serine proteases into the target cell. This is a calcium-dependent process. In addition, FasL expression is upregulated and interaction with Fas can lead to calcium-independent signaling. Both pathways can lead to apoptosis and cell death. In general perforin-mediated lysis is more targeted and FasL-mediated lysis is less specific and more likely to result in antigen-independent bystander killing. We found that coexpression of IL-4 with RSV M2 in recombinant vaccinia increases FasL expression on the M2-specific CD8 ⫹ T cells resulting in diminished perforin-mediated cytolytic activity and enhanced FasL-mediated target cell killing (Aung and Graham, 2000). Experiments in perforin KO mice confirmed that perforin was not necessary for RSV clearance, but when other cytolytic mechanisms are used, clearance and disease expression are delayed, and illness is more prolonged. In addition, excess production of IFN-␥ and TNF-␣ in lung temporally correlated with the illness (Aung et al., 2001). In contrast, when FasL-defective (gld) mice are infected with RSV, viral clearance occurs with normal kinetics with minimal illness and normal amounts of IFN-␥ and TNF-␣ (J. A. Rutigliano, unpublished observations). These observations suggest a new working hypothesis for the mechanism of severe illness in RSV. We propose that the critical pathogenic event that causes illness in primary RSV infection is an aberrant CD8 ⫹ T cell response. Specifically, inefficient cytolytic mechanisms will lead to delayed and prolonged virus clearance, and bystander injury may be the basis of enhanced pathology in settings where too much reliance is placed on FasL-mediated mechanisms of lysis and cytotoxic cytokines (Fig. 3). An important corollary of this hypothesis is that immunization approaches that induce subsets of CD8 ⫹ CTL responses with efficient, targeted killing mechanisms should result in more favorable outcomes. These concepts will be the subject of future studies in murine models and clinical trials. REFERENCES Anderson, L. J., Tsou, C., Potter, C., Keyserling, H. L., Smith, T. F., Ananaba, G., and Bangham, C. R. (1994). Cytokine response to respiratory syncytial virus stimulation of human peripheral blood mononuclear cells. J. Infect. Dis. 170, 1201–1208. Aung, S., and Graham, B. S. (2000). Differential regulation of perforinand FasL-mediated cytotoxicity by IL-4. J. Immunol. 164, 3487–3493. Aung, S., Rutigliano, J. A., and Graham, B. S. (2001). Alternative mechanisms of respiratory syncytial virus clearance in perforin knockout mice lead to enhanced disease. J. Virol. 75, 9918–9924. Aung, S., Tang, Y. W., and Graham, B. S. (1999). IL-4 inhibits induction of cytotoxic T lymphocyte activity in mice infected with recombinant
vaccinia virus expressing respiratory syncytial virus M2 protein. J. Virol. 73, 8944–8949. Barr, F. E., Pedigo, H., Johnson, T. R., and Shepherd, V. L. (2000). Surfactant protein-A enhances uptake of respiratory syncytial virus by monocytes and U937 macrophages. Am. J. Respir. Cell Mol. Biol. 23, 586–592. Bembridge, G. P., Lopez, J. A., Bustos, R., Melero, J. A., Cook, R., Mason, H., and Taylor, G. (1999). Priming with a secreted form of the fusion protein of respiratory syncytial virus (RSV) promotes interleukin-4 (IL-4) and IL-5 production but not pulmonary eosinophilia following RSV challenge. J. Virol. 73, 10086–10094. Brandenburg, A. H., Kleinjan, A., van Het Land, B., Moll, H. A., Timmerman, H. H., de Swart, R. L., Neijens, H. J., Fokkens, W., and Osterhaus, A. D. (2000). Type 1-like immune response is found in children with respiratory syncytial virus infection regardless of clinical severity. J. Med. Virol. 62, 267–277. Chin, J., Magoffin, R. L., Shearer, L. A., Schieble, J. H., and Lennette, E. H. (1969). Field evaluation of a respiratory syncytial virus vaccine and a trivalent parainfluenza virus vaccine in a pediatric population. Am. J. Epidemiol. 89, 449–463. Domachowske, J. B., Bonville, C. A., Dyer, K. D., Easton, A. J., and Rosenberg, H. F. (2000). Pulmonary eosinophilia and production of MIP-1alpha are prominent responses to infection with pneumonia virus of mice. Cell Immunol. 200, 98–104. Durbin, J., Johnson, T. R., Durbin, R. K., Mertz, S., Peebles, R. S., Jr., and Graham, B. S. (2002). The role of interferon in RSV pathogenesis. J. Immunol. 168, 2944–2951. Falsey, A. R., Cunningham, C. K., Barker, W. H., Kouides, R. W., Yuen, J. B., Menegus, M., Weiner, L. B., Bonville, C. A., and Betts, R. F. (1995). Respiratory syncytial virus and influenza A infections in the hospitalized elderly. J. Infect. Dis. 172, 389–394. Feldman, S. A., Hendry, R. M., and Beeler, J. A. (1999). Identification of a linear heparin binding domain for human respiratory syncytial virus attachment glycoprotein G. J. Virol. 73, 6610–6617. Fishaut, M., Tubergen, D., and McIntosh, K. (1980). Cellular response to respiratory viruses with particular reference to children with disorders of cell-mediated immunity. J. Pediatr. 96, 179–186. Fischer, J. E., Johnson, J. E., Kuli-Zade, R., Johnson, T. R., Aung, S., Parker, R. A., and Graham, B. S. (1997). Overexpression of IL-4 delays virus clearance in mice infected with RSV. J. Virol. 71, 8672–8677. Fischer, J. E., Johnson, T. R., Peebles, R. S., and Graham, B. S. (1999). Vaccination with pertussis toxin alters the antibody response to simultaneous respiratory syncytial virus challenge. J. Infect. Dis. 180, 714–771. Garofalo, R., Kimpen, J. L. L., Welliver, R. C., and Ogra, P. L. (1992). Eosinophil degranulation in the respiratory tract during naturally acquired respiratory syncytial virus infection. J. Pediatr. 120, 28–32. Glezen, P., and Denny, F. W. (1973). Epidemiology of acute lower respiratory disease in children. N. Engl. J. Med. 288, 498–505. Graham, B. S., Bunton, L. A., Wright, P. F., and Karzon, D. T. (1991b). Role of T lymphocyte subsets in the pathogenesis of primary infection and rechallenge with respiratory syncytial virus in mice. J. Clin. Invest. 88, 1026–1033. Graham, B. S., Henderson, G. S., Tang, Y. W., Lu, X., Neuzil, K. M., and Colley, D. G. (1993). Priming immunization determines T helper cytokine mRNA expression patterns in lungs of mice challenged with respiratory syncytial virus. J. Immunol. 151, 2032–2040. Hancock, G. E., Speelman, D. J., Heers, K., Bortell, E., Smith, J., and Cosco, C. (1996). Generation of atypical pulmonary inflammatory responses in BALB/c mice after immunization with the native attachment (G) glycoprotein of respiratory syncytial virus. J. Virol. 70, 7783– 7791. Hendricks, D. A., Baradaran, K., McIntosh, K., and Patterson, J. L. (1987). Appearance of a soluble form of the G protein of respiratory syncytial virus in fluids of infected cells. J. Gen. Virol. 68, 1705–1714. Hertz, M. I., Englund, J. A., Snover, D., Bitterman, P. B., and McGlave, P. B. (1989). Respiratory syncytial virus-induced acute lung injury in
MINIREVIEW adult patients with bone marrow transplants: A clinical approach and review of the literature. Medicine 68, 269–281. Hussell, T., Baldwin, C. J., O’Garra, A., and Openshaw, P. J. M. (1997). CD8⫹ T cells control Th2-driven pathology during pulmonary respiratory syncytial virus infection. Eur. J. Immunol. 27, 3341–3349. Hussell, T., Georgiou, A., Sparer, T. E., Matthews, S., Pala, P., and Openshaw, P. J. (1998). Host genetic determinants of vaccine-induced eosinophilia during respiratory syncytial virus infection. J. Immunol. 161, 6215–6222. Hussell, T., and Openshaw, P. J. (1998). Intracellular IFN-gamma expression in natural killer cells precedes lung CD8⫹ T cell recruitment during respiratory syncytial virus infection. J. Gen. Virol. 79, 2593– 2601. Johnson, T. R., and Graham, B. S. (1999). Secreted respiratory syncytial virus (RSV) G protein induces IL-5, IL-13, and eosinophilia by an IL-4-independent mechanism. J. Virol. 73, 8485–8495. Johnson, T. R., Johnson, J. E., Roberts, S. R., Wertz, G. W., Parker, R. A., and Graham, B. S. (1998). Priming with secreted glycoprotein G of respiratory syncytial virus (RSV) augments interleukin-5 production and tissue eosinophilia after RSV challenge. J. Virol. 72, 2871–2880. Johnson, T. R., Parker, R. A., Johnson, J. E., and Graham, B. S. IL-1 is sufficient for respiratory syncytial virus (RSV) G glycoprotein-induced eosinophilia following RSV challenge. J. Immunol., submitted. Kim, H. W., Canchola, J. G., Brandt, C. D., Pyles, G., Chanock, R. M., Jensen, K., and Parrott, R. H. (1969). Respiratory syncytial virus disease in infants despite prior administration of antigenic inactivated vaccine. Am. J. Epidemiol. 89, 422–434. Kim, H. W., Leikin, S. L., Arrobio, J., Brandt, C. D., Chanock, R. M., and Parrott, R. H. (1976). Cell-mediated immunity to respiratory syncytial virus induced by inactivated vaccine or by infection. Pediatr. Res. 10, 75–78. Krinzman, S., Basgoz, N., Kradin, R., Shepard, J. A., Flieder, D. B., Wright, C. D., Wain, J. C., and Ginns, L. C. (1998). Respiratory syncytial virus-associated infections in adult recipients of solid organ transplants. J. Heart Lung Transplant. 17, 202–210. Lambert, D. M. (1988). Role of oligosaccharides in the structure and function of respiratory syncytial virus glycoproteins. Virology 164, 458–466. Murphy, B. R., Prince, G. A., Walsh, E. E., Kim, H. W., Parrott, R. H., Hemming, V. G., Rodriguez, W. J., and Chanock, R. M. (1986). Dissociation between serum neutralizing and glycoprotein antibody responses of infants and children who received inactivated respiratory syncytial virus vaccine. J. Clin. Microbiol. 124, 197–202. Murphy, B. R., and Walsh, E. E. (1988). Formalin-inactivated respiratory syncytial virus vaccine induces antibodies to the fusion glycoprotein that are deficient in fusion-inhibiting activity. J. Clin. Microbiol. 26, 1595–1597. Naniche, D., Yeh, A., Eto, D., Manchester, M., Friedman, R. M., and Oldstone, M. B. (2000). Evasion of host defenses by measles virus: Wild-type measles virus infection interferes with induction of alpha/ beta interferon production. J. Virol. 74, 7478–7484. Openshaw, P. J. M., Clarke, S. L., and Record, F. M. (1992). Pulmonary eosinophilic response to respiratory syncytial virus infection in mice sensitized to the major surface glycoprotein G. Inter. Immunol. 4, 493–500. Peebles, R. S., Sheller, J. R., Collins, R. D., Jarzecka, A. K., Mitchell, D. B., Parker, R. A., and Graham, B. S. (2001a). Respiratory syncytial virus
7
infection decreases allergen-induced type 2 cytokine production while increasing airway hyperresponsiveness in mice. J. Med. Virol. 63, 178–188. Peebles, R. S., Hashimoto, K., Collins, R. D., Jarzecka, K., Furlong, J., Mitchell, D. B., Sheller, J. R., and Graham, B. S. (2001b). Immune interaction between respiratory syncytial virus infection and allergen sensitization critically depends on timing of challenges. J. Infect. Dis. 184, 1374–1379. Roberts, S. R., Lichtenstein, D., Ball, L. A., and Wertz, G. W. (1994). The membrane-associated and secreted forms of the respiratory syncytial virus attachment glycoprotein G are synthesized from alternative initiation codons. J. Virol. 68, 4538–4546. Samore, M. H., and Siber, G. R. (1996). Pertussis toxin enhanced IgG1 and IgE responses to primary tetanus immunization are mediated by interleukin-4 and persist during secondary responses to tetanus alone. Vaccine 14, 290–297. Schlender, J., Bossert, B., Buchholz, U., and Conzelmann, K. K. (2000). Bovine respiratory syncytial virus nonstructural proteins NS1 and NS2 cooperatively antagonize alpha/beta interferon-induced antiviral response. J. Virol. 74, 8234–8242. Sigurs, N., Bjarnason, R., Sigurbergsson, F., Kjellman, B., and Bjorksten, B. (1995). Asthma and immunoglobulin E antibodies after respiratory syncytial virus bronchiolitis: A prospective cohort study with matched controls. Pediatrics 95, 500–505. Sparer, T. E., Matthews, S., Hussell, T., Rae, A. J., Garcia-Barreno, B., Melero, J. A., and Openshaw, P. J. M. (1998). Eliminating a region of respiratory syncytial virus attachment protein allows induction of protective immunity without vaccine-enhanced lung eosinophilia. J. Exp. Med. 187, 1921–1926. Srikiatkachorn, A., and Braciale, T. J. (1997). Virus-specific CD8⫹ T lymphocytes downregulate T helper cell type 2 cytokine secretion and pulmonary eosinophilia during experimental murine respiratory syncytial virus infection. J. Exp. Med. 186, 421–432. Srikiatkhachorn, A., Chang, W., and Braciale, T. J. (1999). Induction of Th-1 and Th-2 responses by respiratory syncytial virus attachment glycoprotein is epitope and major histocompatibility complex independent. J. Virol. 73, 6590–6597. Tang, Y. W., and Graham, B. S. (1994). Anti-IL-4 treatment at immunization modulates cytokine expression, reduces illness, and increases cytotoxic T lymphocyte activity in mice challenged with respiratory syncytial virus. J. Clin. Invest. 94, 1953–1958. Tebbey, P. W., Hagen, M., and Hancock, G. E. (1998). Atypical pulmonary eosinophilia is mediated by a specific amino acid sequence of the attachment (G) protein of respiratory syncytial virus. J. Exp. Med. 188, 1967–1972. Tripp, R. A., Jones, L. P., Haynes, L. M., Zheng, H., Murphy, P. M., and Anderson, L. J. (2001). CX3C chemokine mimicry by respiratory syncytial virus G glycoprotein. Nat. Immunol. 2, 732–738. Varga, S. M., Wang, X., Welsh, R. M., and Braciale, T. J. (2001). Immunopathology in RSV infection is mediated by a discrete oligoclonal subset of antigen-specific CD4(⫹) T cells. Immunity 15, 637–646. Welliver, R. C., Wong, D. T., Sun, M., Middleton, E., Jr., Vaughan, R. S., and Ogra, P. L. (1981). The development of respiratory syncytial virus-specific IgE and the release of histamine in nasopharyngeal secretions after infection. N. Engl. J. Med. 305, 841–884. Wills-Karp, M. (1999). Immunologic basis of antigen-induced airway hyperresponsiveness. Annu. Rev. Immunol. 17, 255–281.