Friday 11 November 2011
Mechanisms of Respiratory Syncytial Virus speciic T cell activation
Promotor: Prof. dr J.L. Kimpen
Defence: 11 November 2011
During a natural infection with RSV mainly respiratory epithelial cells and antigen presenting cells, like alveolar macrophages and dendritic cells (DC) become infected. Infection of these cells results in initiation of innate immune responses, including the production of interferons. The production of interferons is important for inhibition of viral replication and activation of immune cells. In addition, the production of chemokines and cytokines by innate immune cells triggers an influx of inflammatory cells into the lung. Initiation of the adaptive immune response is predominantly mediated via DCs, which are the professional antigen presenting cells. DCs capture antigen either via direct infection or via uptake of viral material through the endosomal route. These cells transport viral material to the lung draining lymph node and present antigenic peptides to CD4+ and CD8+ T cells. Activated T cells migrate back to the site of inflammation to eliminate virus infected cells. We show in chapter 2 that different subsets of lung DCs present antigenic epitopes of RSV in the context of MHC class I and class II molecules in mice. The CD103+CD11b- DCs located directly underneath the airway epithelia and the CD103-CD11b+ DCs from the lung parenchyma, become infected with RSV. Upon migration to the lung draining lymph nodes, these DCs present antigens to CD4+ and CD8+ T cells. The activated T cells migrate back to the site of inflammation where they eliminate virus infected cells. Furthermore, also a subset of lymph node resident DCs acquire viral antigens via a non-infectious route and these cells are also capable to activate RSV specific effector/memory CD4+ and CD8+ T cells. It remains to be established whether all these DC types are also able to activate naïve T cells in the lymph nodes. However, to test this T cell receptor transgenic mice are required that are currently not available. After an initial wave of DC migration to the lung draining lymph nodes, the lungs are repopulated with DCs most likely derived from blood monocytes or DC precursors in the lungs.
The activation of the adaptive immune response as described above might be different during primary RSV infections and secondary infections, i.e. in the presence of pre-existing immunity like RSV specific T cells and virus specific antibodies. In mouse and ex vivo human studies we addressed the question how antibodies affect (primary) RSV infections and how they modulate the initiation of adaptive immune responses. Antibodies specific for RSV might be obtained by the infant from the mother via the placenta and breast milk or acquired after natural infection. During viral infections antibodies can have different effects on the immune response. Neutralizing antibodies lower the viral load and thereby reduce pathogenesis induced by viral infection. By lowering viral load antibodies can also reduce innate immune responses and inflammation or alter innate immune responses by targeting pathogens to intracellular compartments, where different TLRs are present compared to the cell surface. Both neutralizing and non-neutralizing antibodies can play a role in the antigen presentation process and determine the level of CD8+ and CD4+ T cell activation.
We studied the effect of antibodies on the initiation of RSV specific T cell responses in a mouse RSV infection model using wild type C57BL/6 and FcR knockout mice. In chapter 3 we showed that antibodies critically affected the RSV specific CD4+/CD8+ T cell balance in the mouse models. We had found earlier that during a primary RSV infection in children admitted to the intensive care unit mainly virus specific CD8+ T cell responses were observed in peripheral blood and hardly any CD4+ T cells. In contrast, in healthy adults very low numbers of RSV specific CD8+ T cells were detectable in peripheral blood while CD4+ T cells were more abundant (chapter 3). We showed in vitro using PBMCs that non-neutralizing and neutralizing antibodies enhanced RSV specific CD4+ T cell responses while neutralizing antibodies decreased CD8+ T cell responses. These different effects of neutralizing and non-neutralizing antibodies were also observed in the in vivo mouse model. From this observation that neutralizing antibodies decreased the balance of CD8+/CD4+ virus specific T cells it appears that antibody mediated cross-presentation was presumably a minor route by which CD8+ T cell responses were elicited, while direct infection of APC or possibly cross-presentation of dying epithelial cells might be more important.
The altered balance between RSV specific CD4+ and CD8+ T cell responses in the presence of antibodies suggested an altered route for antigen presentation of RSV opsonized by IgGs. FcγRs are expressed on antigen presenting cells, they bind immune complexes (IC) and promote internalization of the antigen into the endosomal route. Internalization of IC into the endosomal route facilitates MHC class II antigen presentation and subsequent CD4+ T cell activation. In wild type mice we showed that RSV-IC were more efficiently presented to CD4+ T cells when activating FcγRs were expressed compared to the sitiuation in FcγR knockout mice, in both experiments in vivo and in vitro (chapter 3). The inhibiting FcγR-IIb did not play such a role. In conclusion, both neutralizing and non-neutralizing antibodies increased the CD4+ T cell response via an FcγR dependent antigen presentation route, while neutralizing antibodies lowered CD8+ T cell responses. It is interesting to speculate that the existing neutralizing antibody response in RSV-immune individuals might favor CD4+ T cell expansion and thus explain the low virus CD8+/CD4+ T cell ratio for RSV in adults and elderly individuals. In contrast to the situation for RSV, influenza virus CD8+ T cells outnumber CD4+ T cells. The difference might be explained by the fact that influenza viruses escape antibody neutralization and therefore efficiently boost CD8+ T cell responses in immune individuals.
In addition to FcγRs, FcRn binds the IgG Fc domain, but at a different site then FcγR. Due to a high binding affinity for IgG at low pH FcRn binds IgG in endosomes, whereas FcγRs bind IgG at the cell surface at neutral pH. FcRn plays a role in protecting antibodies from degradation in lysosomes by shuttling endocytosed IgGs back to the cell surface, where they are released. In addition to its function in increasing antibody half life, FcRn plays a role in antibody translocation across epithelial barriers. FcRn is located in gut and lung epithelium and facilitates antibody translocation into gut and airways and vice versa. FcRn is also important in translocation of IgG from mother to child across the placenta. Recently, FcRn was found to mediate MHC class II mediated antigen presentation of the model antigen OVA, when it was opsonized by IgGs. Since we found that FcγRs played an important role in the antigen presentation in immune mice, we decided to study the role of FcRn in antigen presentation during RSV infection. We showed for RSV that FcRn was not involved in antigen presentation to CD4+ T cells (chapter 4). Furthermore, we tested whether FcRn was involved in immune complex transport across the lung epithelial barrier during RSV infections facilitating viral antigen access to APC in the lungs. We showed indeed that increased T cell responses were found when RSV was applied as an immune complex with neutralizing antibodies compared to inoculation of UV-inactivated RSV. However, this enhanced response was FcRn independent (chapter 4).
Despite the fact that a (primary) RSV infection results in the induction of CD4+ and CD8+ T cell responses and stimulates the production of neutralizing antibodies, re-infections occur frequently even with genetically similar virus strains. The fact that natural immunity is not 100% effective against re-infections is a hurdle for vaccine development. It is important to understand how highly neutralizing antibodies and efficient T cell responses can be induced and maintained via a primary exposure to RSV or an RSV vaccine. Several vaccination approaches have been studied such as formalin-inactivated RSV (FI-RSV), life-attenuated viruses and subunit (protein) vaccines. A clinical trial with FI-RSV performed in the 1960s resulted in a disaster. Children who received FI-RSV intramuscularly and experienced a natural RSV infection, developed severe lower respiratory tract disease compared to control formalin-inactivated parainfluenza vaccine recipients.
Different animal models were used to study the FI-RSV mediated enhanced disease. In BALB/c mice FI-RSV vaccination results in a strong Th2 response with a central role for CD4+ T cells and pulmonary influx of eosinophils/neutrophils. This is accompanied by decreased CD8+ T cell responses, poor neutralizing antibody responses, airway hypersensitivity, increased mucus production and weight loss. We performed FI-RSV vaccination experiments in C57BL/6 mice (chapter 5) to understand the role of different components of the vaccine contributing to the harmful immune response. We found CD4+/Th2 cells dominating the immune responses, similar to the BALB/c model. We showed in our study that contrary to earlier reports the typical Th2/eosinophilic response was independent of formalin-inactivation of the virus (epitope disruption or carbonyl mediated adjuvant effects). Also the adjuvant aluminum hydroxide (an adjuvant known to shift the recall response towards Th2 was not crucial. We and others showed in mice that the strong Th2 response mediated by FI-RSV vaccination was similar to the allergic response observed in mice vaccinated with the control vaccine. Both FI-RSV (cultured in FCS) and FCS alone administered i.m. induced a Th2 response upon i.n. RSV/FCS challenge. Thus specific characteristics of RSV proteins were not responsible for the deleterious immune response as reported earlier, where particularly the G protein was claimed to induce the extreme Th2 response.
Inactivated parainfluenza virus, used as the control vaccine in the 1960s trial, did not cause serious lower respiratory tract disease. We therefore reasoned that not only the specific mode of vaccination, but also characteristics of the virus might contribute to the particular type of immune response upon natural virus exposure. We indeed showed that intramuscular priming of mice with FCS and an intranasal challenge with influenza virus plus FCS resulted in a reduced Th2 response compared to mice challenged with RSV plus FCS. This suggested that different respiratory viruses can locally influence the nature of T cell responses, presumably by the nature of the local innate response triggered by the infecting virus. We hypothesize that different innate signatures of natural respiratory infections could possibly also explain different effects during allergic/asthmatic responses.
In our study described in chapter 5 we additionally showed that T cell responses in the lung could be manipulated by the way the immune response was primed. Priming of a Th1 skewed CD4+ T cell response or a CD8+ T cell response (in chapter 5 accomplished by i.v. injection of matured antigen loaded dendritic cells) resulted in decreased Th2 immunity locally in the lungs of RSV/FCS challenged mice.
Thus there are three critical elements in inactivated RSV vaccines that contribute to the characteristic unfavorable immune response; 1. the absence of priming of a CD8+ or Th1-skewed CD4+ T cell response by inactivated RSV, 2. the innate immune response induced by a natural RSV infection in the lungs is inadequate to shift the Th0, Th2 response induced during priming towards a more protective Th1 response, or the innate response creates a situation whereby Th2 cells are preferably attracted, 3. the virus specific antibodies produced are poorly neutralizing, hence not protecting the host against viral damage, but importantly also not efficiently protecting the host against a high antigenic load in the face of an already primed CD4+ T cell response.
Infants who receive breast milk are protected against severe RSV infections. This is partially explained by the neutralizing antibodies present in breast milk that are transferred to the child. However, more compounds present in breast milk might contribute to protection like non-digestible oligosaccharides. These non-digestible oligosaccharides have been attributed immune modulating effects that may be ascribed to their ability to affect the composition of intestinal microbial flora. Recent studies on the effect of these compounds in breast milk to prevent atopic diseases were encouraging. It has been shown that oligosaccharides mimicking those compounds present in breast milk can alter the Th2/Th1 balance. Because one explanation for severe primary RSV infections could be the relative immaturity and Th2 bias of the infant immune system we hypothesized that oligosaccharide mediated immune modulation might affect the RSV specific immune response. We studied the effect of a specific oligosaccharide diet in a FI-RSV induced Th2 response in our murine vaccination model (chapter 6). We indeed observed that mice on a diet with specific non-digestible oligosaccharides showed a reduced percentage of Th2 cytokine producing CD4+ T cells in FI-RSV vaccinated/RSV challenged mice compared to the response in mice that received control diet. The exact mechanism how oligosaccharides influence the immune response needs to be elucidated.