Vaccination with an Acidic Polymerase Epitope of Influenza Virus Elicits a Potent Antiviral T Cell Response but Delayed Clearance of an Infl
http://www.100md.com
免疫学杂志 2005年第2期
Abstract
The mechanisms underlying epitope selection and the potential impact of immunodominance hierarchies on peptide-based vaccines are not well understood. Recently, we have shown that two immunodominant MHC class I-restricted epitopes, NP366–374/Db (nucleoprotein (NP)) and PA224–233/Db (acidic polymerase (PA)), which drive the CD8+ T cell response to influenza virus infection in C57BL/6 mice, are differentially expressed on infected cells. Whereas NP appears to be strongly expressed on all infected cells, PA appears to be strongly expressed on dendritic cells but only weakly expressed on nondendritic cells. Thus, the immune response to influenza virus may involve T cells specific for epitopes, such as PA, that are poorly expressed at the site of infection. To examine the consequences of differential Ag presentation on peptide vaccination, we compared the kinetics of the T cell response and influenza virus clearance in mice vaccinated with the NP or PA peptide. Vaccination with either the NP or PA peptide resulted in accelerated and enhanced Ag-specific T cell responses at the site of infection following influenza virus challenge. These T cells were fully functional in terms of their ability to produce IFN- and TNF- and to mediate cytolytic activity. Despite this enhancement of the Ag-specific T cell response, PA vaccination had a detrimental effect on the clearance of influenza virus compared with unvaccinated or NP-vaccinated mice. These data suggest that differential Ag presentation impacts the efficacy of T cell responses to specific epitopes and that this needs to be considered for the development of peptide-based vaccination strategies.
Introduction
It has been well established that the CD8+ T cell response to viral infections is typically directed to a relatively limited number of epitopes, a phenomenon referred to as immunodominance (1). These epitopes can be ordered into immunodominance hierarchies with dominant epitopes driving the bulk of the T cell response and subdominant epitopes driving comparatively minor portions of the response. The mechanisms underlying epitope selection during an immune response are not well understood, although Ag availability, Ag processing, epitope stability, and T cell repertoire all play critical roles (1, 2, 3, 4, 5, 6, 7). Immunodominance hierarchies are clearly important for the development of vaccines designed to promote cellular immune responses. For example, several studies have shown that subdominant epitopes can serve as good vaccine targets for the control of respiratory virus infections (8, 9). However, the factors that need to be considered for the selection of epitopes used in subunit vaccines that promote protective cellular immune responses remain unclear.
Recently, Belz et al. (10) demonstrated that T cell immunodominance patterns can differ substantially between primary and secondary responses to influenza virus infection in C57BL/6 mice. Whereas T cells specific for two major epitopes, NP366–374/Db (nucleoprotein (NP))3 and PA224–233/Db (acidic polymerase (PA)), are present in equivalent numbers in the lung during the primary response to influenza virus infection, T cells specific for the NP366–374/Db epitope dominate the secondary response (10). We have shown that the NP366–374/Db and PA224–233/Db epitopes of influenza virus are differentially presented by dendritic and nondendritic cells, such as lung epithelial cells, following infection and have speculated that this regulates the specificity of the secondary response (11). The PA224–233/Db epitope appears to be presented most strongly by dendritic cells and only weakly on nondendritic cells, whereas the NP366–374/Db epitope appears to be strongly expressed by many cell types. We believe that more widespread expression of the NP366–374/Db epitope may favor the expansion of NP366–374/Db-specific memory CD8+ T cells during secondary infection, potentially explaining the changing pattern of immunodominance in this system (11).
An interesting feature of this system is that the PA224–233/Db epitope appears to be poorly expressed in cells such as lung epithelial cells that are primary targets of influenza virus infection. This raised the question whether PA224–233/Db-specific T cells would be inefficient at mediating clearance of an influenza virus infection and whether vaccination with PA224–233 would be ineffective. Indeed, preliminary studies suggested that this may be the case, although the quality of the T cell response and the kinetics of viral clearance were not measured (11). In this report, we undertook a detailed analysis of the consequences of PA224–233 peptide vaccination on antiviral immunity. The data show that vaccination with dendritic cells pulsed with either the NP366–374 and PA224–233 peptide of influenza resulted in increased epitope-specific CTLs capable of migrating to the lung and producing IFN- and TNF- after infection. However, despite the strong PA224–233/Db-specific T cell response, PA224–233 vaccination did not accelerate viral clearance.
Materials and Methods
Viruses, animals, and infections
The influenza virus A/HK-x31 (x31, H3N2) was grown, stored, and titrated as previously described (12). Female C57BL/6 mice were purchased from The Jackson Laboratory. Mice (6–12 wk) were anesthetized by i.p. injection with 2,2,2-tribromoethanol and infected intranasally with 300 50% egg infectious doses (EID50) of x31 influenza virus.
Bone marrow-derived dendritic cells
Bone marrow was flushed from the femurs of C57BL/6 mice, depleted of erythrocytes, and 2 x 106 mononuclear cells were placed into a bacteriological Petri dish with medium supplemented with 20 ng/ml recombinant murine GM-CSF (PeproTech) and incubated at 37°C with 10% CO2 (13). On day 3, an additional 10 ml of complete tumor medium (CTM) containing 20 ng/ml recombinant murine GM-CSF was added. On days 6 and 8, half of the cells was removed, centrifuged, and added back to the same plate in 10 ml of fresh medium containing 20 ng/ml rmGM-CSF. On day 10 of the culture, the cells were resuspended at 5 x 106/ml and incubated at 37°C for 3 h with peptide at a concentration of 50 μg/ml (14). Following peptide pulsing, the dendritic cells were washed and 100 μl of cells in PBS were injected i.v. into mice at a final concentration of 1 x 106 cells per mouse.
Tissue preparation
Lymphocytes were collected from the bronchoalveolar lavage (BAL), spleens, lung, and mediastinal lymph nodes (MLN) as previously described (15). Following collection, the cells were washed and depleted of erythrocytes. Cells from the BAL were incubated on plastic for 2 h at 37°C to remove adherent cells. Lymphocytes were enriched from the lung tissue by isolation over an 80/40% Percoll gradient and from the MLN and spleen by panning on IgG-coated plates as previously described (16).
MHC tetramer regents and analysis
MHC class I peptide tetramers (NP366–374/Db and PA224–233/Db) were generated by the Molecular Biology Core Facility at the Trudeau Institute as described previously (17). Tetramer staining was performed for 1 h at room temperature, followed by incubation with anti-CD8 Ab conjugated to PerCP (BD Pharmingen), and 200,000 events were collected on a BD Biosciences FACSCalibur flow cytometer. Data were analyzed using FlowJo (Tree Star) software.
Intracellular cytokine staining
Cells isolated from the BAL, lung tissue, MLN, and spleen were cultured at 37°C for 5 h in the presence of 10 μg of the indicated peptide in 250 μl of CTM containing 10 μg/ml brefeldin A (Epicenter Technologies) and 10 U/ml IL-2 (R&D Systems) (14). After culture, the cells were blocked with mAbs to FcRIII/IIR (BD Pharmingen) and stained with anti-CD8 Ab conjugated to PerCP (BD Pharmingen) in PBS/brefeldin A. The cells were then fixed in 2% formaldehyde, permeabilized with buffer containing 0.5% saponin, and stained with anti-IFN- conjugated to PE (BD Pharmingen) and anti-TNF- conjugated to allophycocyanin (BD Pharmingen) mAbs. A total of 200,000 events was collected as described above.
ELISPOT assay
The number of IFN--secreting cells derived from spleens of vaccinated mice was determined after stimulation with Flu-NP366–374, Flu-PA224–233, or Flu-PB1703–711 peptides (New England Peptide) in a standard ELISPOT assay (16). Briefly, 96-well multiscreen HA nitrocellulose plates (Millipore) were coated overnight at 4°C with a 100-μl well of rat anti-mouse IFN- (BD Pharmingen) at a concentration of 10 μg/ml. The plates were then washed and blocked before the addition of titered numbers of CD8-enriched responding cells, irradiated (3000 rad) syngeneic normal spleen cells, 10 μg/ml peptide, and 40 U/ml IL-2. Plates were then incubated 48 h at 37°C and developed overnight with a biotinylated detection Ab, rat anti-mouse IFN- (BD Pharmingen). The plates were then incubated with streptavidin-alkaline phosphatase (DakoCytomation) for 1 h, washed, and incubated with 5-bromo-4-chloro-3-indolyl phosphate/NBT alkaline phosphatase substrate (Sigma-Aldrich) for 2 h at room temperature. Visible spots of IFN--secreting cells were then enumerated using an Olympus SZH stereo zoom microscope system.
Viral titers
lungs were serially diluted and injected into three 10-day-old embryonated hen eggs per sample. After incubation at 35°C for 48 h, allantoic fluid from each egg was sampled and assayed for hemagglutinating activity using chicken RBC as previously described (18). Samples were scored as positive when at least two of the three eggs contained hemagglutinating activity (18).
Cytotoxicity assay
The cytotoxicity activity of BAL cells was determined using 51Cr-labeled untreated NP366–374-pulsed or PA224–233-pulsed target cells as previously described (14). Briefly, target cells (106 EL4 cells) were incubated with 100 μCi of 51Cr (Na51CrO4; Amersham Biosciences) in CTM for 1 h at 37°C. Seventy-five micrograms of peptide was then added, and the cells were incubated for an additional 3 h. BAL cells were collected from either vaccinated or unvaccinated mice on day 10 postinfection with 300 EID50 x31. The cells were incubated for 2 h at 37°C to remove adherent cells and enrich for lymphocytes. The enriched BAL cells were then incubated for 5 h with the target cells and the percentage of specific release was determined by the formula: percentage of specific release = (100 x (experimental – spontaneous)/(maximum – spontaneous)). To determine spontaneous and maximum release, the cells were incubated with either medium alone (spontaneous) or 1% Triton X-100 (maximum).
Results
Kinetics of viral clearance following NP366–374-specific or PA224–233-specific vaccination
To examine the impact of NP and PA peptide priming on the immune response to influenza virus, we vaccinated mice with either NP366–374- or PA224–233-pulsed dendritic cells. Fourteen days later, IFN- ELISPOT analysis demonstrated increased frequencies of epitope-specific CD8+ T cells in the spleen and confirmed successful vaccination of the mice (Fig. 1). Vaccinated mice were then intranasally infected with 300 EID50 x31, and viral titers were determined in the lungs on days 4, 6, 8, and 10 postinfection using an embryonated hen egg assay. Unvaccinated mice had high viral titers on days 4 and 6, but viral titers were decreased by day 8, and only 2 of 10 mice had detectable virus at day 10 (Fig. 2). In contrast, mice that were vaccinated with NP366–374-pulsed dendritic cells (Fig. 2) were already clearing virus by day 6, and the majority of mice had cleared the virus by day 8. Interestingly, mice that were vaccinated with the PA224–233-pulsed dendritic cells (Fig. 2) had high viral titers at days 4 and 6, and the majority of mice had still not cleared virus by day 10. The mice that were vaccinated with the PA224–233 peptide eventually cleared the infection because no virus was detected in the lungs at day 15 postinfection (data not shown). Thus, whereas NP366–374 vaccination resulted in decreased viral titers and slightly accelerated viral clearance compared with unvaccinated mice, PA224–233 vaccination resulted in delayed viral clearance.
FIGURE 1. Vaccination results in increased Ag-specific T cell frequencies in the spleen. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection. Spleens were removed on day 14 postvaccination, enriched for CD8+ cells, incubated for 48 h with irradiated (3000 rad) syngeneic normal spleen cells, peptide, and IL-2 in a standard ELISPOT assay. Shown is the number of IFN--positive cells following incubation with the indicated peptide from one of three representative experiments.
FIGURE 2. Time course of viral clearance in the lungs of vaccinated and control mice. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. The lungs were removed at the indicated time points, homogenized, and used in an egg infectious assay to determine viral titers. Shown is the log viral titer present in the lungs at the indicated time points postinfection. Each symbol represents an individual mouse, and the line represents the average at each time point from two independent experiments (days 4, 6, and 8) and three independent experiments (day 10).
Kinetics of T cell activation and migration following infection of vaccinated mice
We next asked whether the delayed viral clearance in the PA224–233-vaccinated group was caused by decreased activation, decreased migration of activated cells to the site of infection, and/or decreased functional activity. Vaccinated mice were infected with 300 EID50 x31 intranasally, and the number of NP366–374/Db- or PA224–233/Db-specific T cells was determined in the lung airways on days 4, 6, 8, and 10 postinfection. In unvaccinated mice, CD8+ T cells appeared in the lung airways at day 8 postinfection, and the numbers of NP366–374/Db- or PA224–233/Db-specific T cells were roughly equivalent at day 10 (Fig. 3). In the NP366–374-vaccinated mice, the CD8+ T cell response was detected earlier than in the unvaccinated mice, and increased numbers of NP366–374/Db-specific T cells were detected relative to PA224–233/Db-specific T cells. Although there was an increased number of PA224–233/Db cells in the NP366–374-vaccinated mice at day 8 postinfection, this increase was not significant across multiple experiments. Conversely, the PA224–233 vaccination resulted in significantly increased numbers of PA224–233/Db-specific T cells relative to NP366–374/Db-specific T cells (Fig. 3). When compared with unvaccinated mice, the PA224–233/Db-specific T cell response was very strong in the PA224–233-vaccinated mice, with >5-fold higher numbers of Ag-specific cells in the lung airways on day 10.
FIGURE 3. Kinetics of NP366–374/Db- and PA224–233/Db-specific CD8+ T cells in the BAL following infection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. On days 4, 6, 8, and 10, the BAL was collected and stained with anti-CD8 and the NP366–374/Db- or PA224–233/Db-specific tetramers. Shown are the number of NP366–374/Db- (?) and PA224–233/Db- () specific T cells. The data shown are the average and SD of three independent experiments.
Table I shows the absolute number of NP366–374/Db- or PA224–233/Db-specific T cells as well as the NP/PA ratio in the lung airways, lung parenchyma, draining lymph nodes, and spleen on day 10 postinfection. As observed in the lung airways, unvaccinated mice had roughly equivalent numbers of NP366–374/Db- or PA224–233/Db-specific T cells in other tissues, whereas NP366–374 vaccination resulted in increased numbers of NP366–374/Db-specific T cells, and PA224–233 vaccination resulted in increased numbers of PA224–233/Db-specific T cells.
Table I. Number of Ag-specific CD8+ T cells in the BAL, lung, MLN, and spleen on day 10 postinfection
It should be noted from the data in Table I that vaccination with either the NP366–374 or PA224–233 peptide resulted in a compensatory decrease in the number of T cells specific for the reciprocal epitope in the airways on day 10 postinfection. Specifically, upon vaccination with the NP366–374 peptide, there was a decrease in the number of PA224–233/Db-specific T cells compared with the number found in unvaccinated mice (12,665 cells vs 35,375 cells). Similarly, PA224–233 vaccination resulted in decreased numbers of NP366–374/Db-specific T cells compared with unvaccinated mice (6,836 cells vs 30,167 cells). It is also interesting that the NP/PA ratio is skewed in the lungs compared with the MLN and spleen of PA224–233-vaccinated mice (Table I). It is possible that comparatively low levels of PA224–233 Ag present in the lung airways results in reduced apoptosis and the accumulation of PA224–233/Db-specific T cells at this site. In contrast, NP366–374/Db-specific T cells may be triggered by high levels of Ag in the lung airways to mediate effector functions and undergo rapid apoptosis. This would result in the observed skewing of the NP/PA ratio at the site of infection. Taken together, the poor viral clearance observed in PA224–233-vaccinated mice could not be explained by low numbers of Ag-specific T cells generated following infection or a reduced capacity to migrate to the site of infection.
Dendritic cell vaccination results in functionally active CD8+ T cells
The data thus far indicate that PA224–233-vaccinated mice generate greatly increased numbers of PA224–233/Db-specific T cells to influenza virus infection, but mediate delayed viral clearance. Thus, we investigated the functional properties of the NP366–374/Db- and PA224–233/Db-specific T cells generated under these conditions. As shown in Fig. 4, both NP366–374/Db- and PA224–233/Db-specific T cells produced IFN- and TNF- on day 10 postinfection as determined by an intracellular cytokine-staining assay. When calculated as absolute numbers of cells, the number of IFN--producing cells matched the numbers of cells determined by tetramer analysis. In contrast, only a subset of the IFN--producing cells produced TNF- in response to peptide stimulation. Interestingly, a much higher frequency of PA224–233/Db-specific T cells produced both IFN- and TNF- compared with NP366–374/Db-specific T cells (Fig. 4B) in response to the PA224–233 peptide than to the NP366–374 peptide (Fig. 4B). Whereas 30% of the NP366–374/Db-specific T cells produced TNF-, 75% of the PA224–233/Db-specific T cells produced TNF-. This is consistent with findings by others (19, 20, 21) that PA224–233/Db-specific T cells produce more TNF- upon restimulation than NP366–374/Db-specific T cells.
FIGURE 4. IFN- and TNF- production by Ag-specific CD8+ T cells on day 10 postinfection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31. On day 10 postinfection, the lungs were collected, enriched for lymphocytes, and incubated for 5 h in the presence of the indicated peptides and brefeldin A. Following stimulation, the cells were stained with anti-CD8 PerCP, anti-IFN- PE, and anti-TNF- allophycocyanin. A, The percentages of CD8 cells producing TNF- and IFN- following stimulation with either the NP366–374 or PA224–233 peptide. B, The number NP366–374/Db- and PA224–233/Db- specific cells as determined by tetramer staining, IFN- production, or TNF- production from one of three representative experiments.
In addition to determining the ability of T cells generated by vaccination to produce cytokines, we also determined their cytolytic activity. As seen in Fig. 5, following NP366–374 vaccination there were significantly higher numbers of NP366–374/Db-specific CTLs than PA224–233/Db-specific CTLs. Conversely, following PA224–233 vaccination there were significantly increased numbers of PA224–233/Db-specific CTLs compared with the number of NP366–374/Db-specific CTLs.
FIGURE 5. Ag-specific cytotoxicity on day 10 postinfection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. On day 10 postinfection, the BAL was collected, enriched for lymphocytes, and incubated for 5 h with peptide-pulsed target cells labeled with 51Cr. Following incubation, the supernatants were collected and the percentage of specific release was determined. Shown is the percentage of specific release in NP366–374-vaccinated mice and PA224–233-vaccinated mice from one of three representative experiments.
Taken together, the data presented in this study show that vaccination with either NP366–374 or PA224–233 peptide-pulsed dendritic cells results in increased numbers of CD8+ T cells in the lungs following infection. These epitope-specific T cells were capable of producing IFN- and TNF- upon restimulation with Ag and had potent CTL activity. However, vaccination with the NP366–374 peptide resulted in accelerated viral clearance, whereas PA224–233 vaccination resulted in delayed viral clearance.
Discussion
We have investigated the impact of differential Ag presentation upon vaccine efficacy. As shown by others (9, 22, 23), we found that NP366–374 vaccination results in increased epitope-specific T cell numbers, increased peptide-specific IFN- and TNF- production, potent CTL activity, and accelerated viral clearance compared with unvaccinated mice. Conversely, we found that although PA224–233 vaccination resulted in increased numbers of epitope-specific T cells that are capable of migrating to the site of infection, producing IFN- and TNF- in response to Ag, and elaborating potent CTL activity, this response resulted in delayed viral clearance compared with that of unvaccinated mice. Together, these data support the idea that PA224–233/Db-specific T cells generated following vaccination are fully functional but may be inefficient at clearing virus due to poor expression of the PA224–233 epitope in the lungs. Importantly, the fact that PA vaccination actually resulted in delayed viral clearance compared with unvaccinated mice demonstrates that some epitopes may be detrimental in the context of a vaccine.
It is currently unclear why PA224–233-vaccinated mice exhibit a delay in viral clearance. Based on the data in Fig. 3, it appears likely that the enhanced PA224–233/Db-specific T cell response actually suppressed the NP366–374/Db-specific T cell response. Thus, PA224–233 vaccination may not only elicit T cells that do not mediate protection, but also decrease the production of protective NP366–374/Db-specific T cells (and probably protective T cells of other specificities). However, it should be noted that PA224–233-vaccinated mice were ultimately able to clear the viral infection (data not shown), presumably after the appearance of sufficient numbers of protective T cells specific for other influenza virus epitopes, such as NP366–374 and PB1703–711 (9, 22, 23). In this regard, the number of NP366–374/Db-specific T cells in the airways and lungs are roughly equivalent in unvaccinated and PA224–233-vaccinated mice by day 15 postinfection (data not shown).
One possible explanation for the reduced capacity of PA224–233/Db-specific T cells to clear virus is that they generate an inappropriate antiviral response. Interestingly, we found that although the majority (75%) of IFN--producing PA224–233/Db-specific T cells also produce TNF-, only 30% of the NP366–374/Db-specific T cells produce both IFN- and TNF-. Recent reports by Turner et al. (21) and La Gruta et al. (20) show that following infection of normal C57BL/6 mice, the PA224–233/Db-specific T cell population makes more IFN- and TNF- following short-term in vitro restimulation than the NP366–374/Db-specific T cell population due to higher functional avidity of Ag binding by PA224–233/Db-specific TCR. This difference in TNF- production results in a selective TNFR2-mediated deletion of high avidity PA224–233/Db-specific T cells following in vitro restimulation with high doses of Ag (21). However, this deletion is less apparent in the T cell populations recovered from the lungs of influenza virus-infected mice (21). In this regard, we find that rather than TNF--mediated deletion of PA224–233/Db-specific T cells following infection of vaccinated mice, there is a significant increase in the number of PA224–233/Db-specific in the lungs. Additionally, although this vaccination strategy results in large numbers of the high avidity PA224–233/Db-specific T cells, there is delayed viral clearance. This correlates with our previous data that the lung epithelial cells poorly present the PA224–233/Db epitope, which would result in low Ag levels following infection, survival of the high affinity PA224–233/Db-specific T cells at this site, and delayed viral clearance.
It is currently unclear whether the differential presentation of the PA224–233/Db epitope by dendritic and nondendritic cells is a unique situation or representative of a class of epitopes. Recently, Zhong et al. (24) identified an array of previously unidentified CD8+ T cell epitopes in the PR8/8/34 strain of influenza virus and examined the ability of epitope-specific CD8 T cells to lyse peptide-pulsed or virally infected EL4 target cells. Interestingly, the study identified two new epitopes that, like the PA224–233 epitope, do not appear to be presented by virally infected EL4 cells. Thus, there is an indication that differential Ag processing and presentation may occur with other influenza virus epitopes.
Taken together, the data in this report show that dendritic cell vaccination resulted in increased numbers of epitope-specific CTLs capable of migrating to the site of infection and producing IFN- and TNF-. Additionally, we have shown that vaccination with an epitope presented predominantly by infected dendritic cells, but not lung epithelial cells, resulted in delayed control of a subsequent viral infection. Understanding the impact of differential Ag processing is therefore essential for the development of effective peptide-based vaccines.(Sherry R. Crowe, Shannon )
The mechanisms underlying epitope selection and the potential impact of immunodominance hierarchies on peptide-based vaccines are not well understood. Recently, we have shown that two immunodominant MHC class I-restricted epitopes, NP366–374/Db (nucleoprotein (NP)) and PA224–233/Db (acidic polymerase (PA)), which drive the CD8+ T cell response to influenza virus infection in C57BL/6 mice, are differentially expressed on infected cells. Whereas NP appears to be strongly expressed on all infected cells, PA appears to be strongly expressed on dendritic cells but only weakly expressed on nondendritic cells. Thus, the immune response to influenza virus may involve T cells specific for epitopes, such as PA, that are poorly expressed at the site of infection. To examine the consequences of differential Ag presentation on peptide vaccination, we compared the kinetics of the T cell response and influenza virus clearance in mice vaccinated with the NP or PA peptide. Vaccination with either the NP or PA peptide resulted in accelerated and enhanced Ag-specific T cell responses at the site of infection following influenza virus challenge. These T cells were fully functional in terms of their ability to produce IFN- and TNF- and to mediate cytolytic activity. Despite this enhancement of the Ag-specific T cell response, PA vaccination had a detrimental effect on the clearance of influenza virus compared with unvaccinated or NP-vaccinated mice. These data suggest that differential Ag presentation impacts the efficacy of T cell responses to specific epitopes and that this needs to be considered for the development of peptide-based vaccination strategies.
Introduction
It has been well established that the CD8+ T cell response to viral infections is typically directed to a relatively limited number of epitopes, a phenomenon referred to as immunodominance (1). These epitopes can be ordered into immunodominance hierarchies with dominant epitopes driving the bulk of the T cell response and subdominant epitopes driving comparatively minor portions of the response. The mechanisms underlying epitope selection during an immune response are not well understood, although Ag availability, Ag processing, epitope stability, and T cell repertoire all play critical roles (1, 2, 3, 4, 5, 6, 7). Immunodominance hierarchies are clearly important for the development of vaccines designed to promote cellular immune responses. For example, several studies have shown that subdominant epitopes can serve as good vaccine targets for the control of respiratory virus infections (8, 9). However, the factors that need to be considered for the selection of epitopes used in subunit vaccines that promote protective cellular immune responses remain unclear.
Recently, Belz et al. (10) demonstrated that T cell immunodominance patterns can differ substantially between primary and secondary responses to influenza virus infection in C57BL/6 mice. Whereas T cells specific for two major epitopes, NP366–374/Db (nucleoprotein (NP))3 and PA224–233/Db (acidic polymerase (PA)), are present in equivalent numbers in the lung during the primary response to influenza virus infection, T cells specific for the NP366–374/Db epitope dominate the secondary response (10). We have shown that the NP366–374/Db and PA224–233/Db epitopes of influenza virus are differentially presented by dendritic and nondendritic cells, such as lung epithelial cells, following infection and have speculated that this regulates the specificity of the secondary response (11). The PA224–233/Db epitope appears to be presented most strongly by dendritic cells and only weakly on nondendritic cells, whereas the NP366–374/Db epitope appears to be strongly expressed by many cell types. We believe that more widespread expression of the NP366–374/Db epitope may favor the expansion of NP366–374/Db-specific memory CD8+ T cells during secondary infection, potentially explaining the changing pattern of immunodominance in this system (11).
An interesting feature of this system is that the PA224–233/Db epitope appears to be poorly expressed in cells such as lung epithelial cells that are primary targets of influenza virus infection. This raised the question whether PA224–233/Db-specific T cells would be inefficient at mediating clearance of an influenza virus infection and whether vaccination with PA224–233 would be ineffective. Indeed, preliminary studies suggested that this may be the case, although the quality of the T cell response and the kinetics of viral clearance were not measured (11). In this report, we undertook a detailed analysis of the consequences of PA224–233 peptide vaccination on antiviral immunity. The data show that vaccination with dendritic cells pulsed with either the NP366–374 and PA224–233 peptide of influenza resulted in increased epitope-specific CTLs capable of migrating to the lung and producing IFN- and TNF- after infection. However, despite the strong PA224–233/Db-specific T cell response, PA224–233 vaccination did not accelerate viral clearance.
Materials and Methods
Viruses, animals, and infections
The influenza virus A/HK-x31 (x31, H3N2) was grown, stored, and titrated as previously described (12). Female C57BL/6 mice were purchased from The Jackson Laboratory. Mice (6–12 wk) were anesthetized by i.p. injection with 2,2,2-tribromoethanol and infected intranasally with 300 50% egg infectious doses (EID50) of x31 influenza virus.
Bone marrow-derived dendritic cells
Bone marrow was flushed from the femurs of C57BL/6 mice, depleted of erythrocytes, and 2 x 106 mononuclear cells were placed into a bacteriological Petri dish with medium supplemented with 20 ng/ml recombinant murine GM-CSF (PeproTech) and incubated at 37°C with 10% CO2 (13). On day 3, an additional 10 ml of complete tumor medium (CTM) containing 20 ng/ml recombinant murine GM-CSF was added. On days 6 and 8, half of the cells was removed, centrifuged, and added back to the same plate in 10 ml of fresh medium containing 20 ng/ml rmGM-CSF. On day 10 of the culture, the cells were resuspended at 5 x 106/ml and incubated at 37°C for 3 h with peptide at a concentration of 50 μg/ml (14). Following peptide pulsing, the dendritic cells were washed and 100 μl of cells in PBS were injected i.v. into mice at a final concentration of 1 x 106 cells per mouse.
Tissue preparation
Lymphocytes were collected from the bronchoalveolar lavage (BAL), spleens, lung, and mediastinal lymph nodes (MLN) as previously described (15). Following collection, the cells were washed and depleted of erythrocytes. Cells from the BAL were incubated on plastic for 2 h at 37°C to remove adherent cells. Lymphocytes were enriched from the lung tissue by isolation over an 80/40% Percoll gradient and from the MLN and spleen by panning on IgG-coated plates as previously described (16).
MHC tetramer regents and analysis
MHC class I peptide tetramers (NP366–374/Db and PA224–233/Db) were generated by the Molecular Biology Core Facility at the Trudeau Institute as described previously (17). Tetramer staining was performed for 1 h at room temperature, followed by incubation with anti-CD8 Ab conjugated to PerCP (BD Pharmingen), and 200,000 events were collected on a BD Biosciences FACSCalibur flow cytometer. Data were analyzed using FlowJo (Tree Star) software.
Intracellular cytokine staining
Cells isolated from the BAL, lung tissue, MLN, and spleen were cultured at 37°C for 5 h in the presence of 10 μg of the indicated peptide in 250 μl of CTM containing 10 μg/ml brefeldin A (Epicenter Technologies) and 10 U/ml IL-2 (R&D Systems) (14). After culture, the cells were blocked with mAbs to FcRIII/IIR (BD Pharmingen) and stained with anti-CD8 Ab conjugated to PerCP (BD Pharmingen) in PBS/brefeldin A. The cells were then fixed in 2% formaldehyde, permeabilized with buffer containing 0.5% saponin, and stained with anti-IFN- conjugated to PE (BD Pharmingen) and anti-TNF- conjugated to allophycocyanin (BD Pharmingen) mAbs. A total of 200,000 events was collected as described above.
ELISPOT assay
The number of IFN--secreting cells derived from spleens of vaccinated mice was determined after stimulation with Flu-NP366–374, Flu-PA224–233, or Flu-PB1703–711 peptides (New England Peptide) in a standard ELISPOT assay (16). Briefly, 96-well multiscreen HA nitrocellulose plates (Millipore) were coated overnight at 4°C with a 100-μl well of rat anti-mouse IFN- (BD Pharmingen) at a concentration of 10 μg/ml. The plates were then washed and blocked before the addition of titered numbers of CD8-enriched responding cells, irradiated (3000 rad) syngeneic normal spleen cells, 10 μg/ml peptide, and 40 U/ml IL-2. Plates were then incubated 48 h at 37°C and developed overnight with a biotinylated detection Ab, rat anti-mouse IFN- (BD Pharmingen). The plates were then incubated with streptavidin-alkaline phosphatase (DakoCytomation) for 1 h, washed, and incubated with 5-bromo-4-chloro-3-indolyl phosphate/NBT alkaline phosphatase substrate (Sigma-Aldrich) for 2 h at room temperature. Visible spots of IFN--secreting cells were then enumerated using an Olympus SZH stereo zoom microscope system.
Viral titers
lungs were serially diluted and injected into three 10-day-old embryonated hen eggs per sample. After incubation at 35°C for 48 h, allantoic fluid from each egg was sampled and assayed for hemagglutinating activity using chicken RBC as previously described (18). Samples were scored as positive when at least two of the three eggs contained hemagglutinating activity (18).
Cytotoxicity assay
The cytotoxicity activity of BAL cells was determined using 51Cr-labeled untreated NP366–374-pulsed or PA224–233-pulsed target cells as previously described (14). Briefly, target cells (106 EL4 cells) were incubated with 100 μCi of 51Cr (Na51CrO4; Amersham Biosciences) in CTM for 1 h at 37°C. Seventy-five micrograms of peptide was then added, and the cells were incubated for an additional 3 h. BAL cells were collected from either vaccinated or unvaccinated mice on day 10 postinfection with 300 EID50 x31. The cells were incubated for 2 h at 37°C to remove adherent cells and enrich for lymphocytes. The enriched BAL cells were then incubated for 5 h with the target cells and the percentage of specific release was determined by the formula: percentage of specific release = (100 x (experimental – spontaneous)/(maximum – spontaneous)). To determine spontaneous and maximum release, the cells were incubated with either medium alone (spontaneous) or 1% Triton X-100 (maximum).
Results
Kinetics of viral clearance following NP366–374-specific or PA224–233-specific vaccination
To examine the impact of NP and PA peptide priming on the immune response to influenza virus, we vaccinated mice with either NP366–374- or PA224–233-pulsed dendritic cells. Fourteen days later, IFN- ELISPOT analysis demonstrated increased frequencies of epitope-specific CD8+ T cells in the spleen and confirmed successful vaccination of the mice (Fig. 1). Vaccinated mice were then intranasally infected with 300 EID50 x31, and viral titers were determined in the lungs on days 4, 6, 8, and 10 postinfection using an embryonated hen egg assay. Unvaccinated mice had high viral titers on days 4 and 6, but viral titers were decreased by day 8, and only 2 of 10 mice had detectable virus at day 10 (Fig. 2). In contrast, mice that were vaccinated with NP366–374-pulsed dendritic cells (Fig. 2) were already clearing virus by day 6, and the majority of mice had cleared the virus by day 8. Interestingly, mice that were vaccinated with the PA224–233-pulsed dendritic cells (Fig. 2) had high viral titers at days 4 and 6, and the majority of mice had still not cleared virus by day 10. The mice that were vaccinated with the PA224–233 peptide eventually cleared the infection because no virus was detected in the lungs at day 15 postinfection (data not shown). Thus, whereas NP366–374 vaccination resulted in decreased viral titers and slightly accelerated viral clearance compared with unvaccinated mice, PA224–233 vaccination resulted in delayed viral clearance.
FIGURE 1. Vaccination results in increased Ag-specific T cell frequencies in the spleen. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection. Spleens were removed on day 14 postvaccination, enriched for CD8+ cells, incubated for 48 h with irradiated (3000 rad) syngeneic normal spleen cells, peptide, and IL-2 in a standard ELISPOT assay. Shown is the number of IFN--positive cells following incubation with the indicated peptide from one of three representative experiments.
FIGURE 2. Time course of viral clearance in the lungs of vaccinated and control mice. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. The lungs were removed at the indicated time points, homogenized, and used in an egg infectious assay to determine viral titers. Shown is the log viral titer present in the lungs at the indicated time points postinfection. Each symbol represents an individual mouse, and the line represents the average at each time point from two independent experiments (days 4, 6, and 8) and three independent experiments (day 10).
Kinetics of T cell activation and migration following infection of vaccinated mice
We next asked whether the delayed viral clearance in the PA224–233-vaccinated group was caused by decreased activation, decreased migration of activated cells to the site of infection, and/or decreased functional activity. Vaccinated mice were infected with 300 EID50 x31 intranasally, and the number of NP366–374/Db- or PA224–233/Db-specific T cells was determined in the lung airways on days 4, 6, 8, and 10 postinfection. In unvaccinated mice, CD8+ T cells appeared in the lung airways at day 8 postinfection, and the numbers of NP366–374/Db- or PA224–233/Db-specific T cells were roughly equivalent at day 10 (Fig. 3). In the NP366–374-vaccinated mice, the CD8+ T cell response was detected earlier than in the unvaccinated mice, and increased numbers of NP366–374/Db-specific T cells were detected relative to PA224–233/Db-specific T cells. Although there was an increased number of PA224–233/Db cells in the NP366–374-vaccinated mice at day 8 postinfection, this increase was not significant across multiple experiments. Conversely, the PA224–233 vaccination resulted in significantly increased numbers of PA224–233/Db-specific T cells relative to NP366–374/Db-specific T cells (Fig. 3). When compared with unvaccinated mice, the PA224–233/Db-specific T cell response was very strong in the PA224–233-vaccinated mice, with >5-fold higher numbers of Ag-specific cells in the lung airways on day 10.
FIGURE 3. Kinetics of NP366–374/Db- and PA224–233/Db-specific CD8+ T cells in the BAL following infection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. On days 4, 6, 8, and 10, the BAL was collected and stained with anti-CD8 and the NP366–374/Db- or PA224–233/Db-specific tetramers. Shown are the number of NP366–374/Db- (?) and PA224–233/Db- () specific T cells. The data shown are the average and SD of three independent experiments.
Table I shows the absolute number of NP366–374/Db- or PA224–233/Db-specific T cells as well as the NP/PA ratio in the lung airways, lung parenchyma, draining lymph nodes, and spleen on day 10 postinfection. As observed in the lung airways, unvaccinated mice had roughly equivalent numbers of NP366–374/Db- or PA224–233/Db-specific T cells in other tissues, whereas NP366–374 vaccination resulted in increased numbers of NP366–374/Db-specific T cells, and PA224–233 vaccination resulted in increased numbers of PA224–233/Db-specific T cells.
Table I. Number of Ag-specific CD8+ T cells in the BAL, lung, MLN, and spleen on day 10 postinfection
It should be noted from the data in Table I that vaccination with either the NP366–374 or PA224–233 peptide resulted in a compensatory decrease in the number of T cells specific for the reciprocal epitope in the airways on day 10 postinfection. Specifically, upon vaccination with the NP366–374 peptide, there was a decrease in the number of PA224–233/Db-specific T cells compared with the number found in unvaccinated mice (12,665 cells vs 35,375 cells). Similarly, PA224–233 vaccination resulted in decreased numbers of NP366–374/Db-specific T cells compared with unvaccinated mice (6,836 cells vs 30,167 cells). It is also interesting that the NP/PA ratio is skewed in the lungs compared with the MLN and spleen of PA224–233-vaccinated mice (Table I). It is possible that comparatively low levels of PA224–233 Ag present in the lung airways results in reduced apoptosis and the accumulation of PA224–233/Db-specific T cells at this site. In contrast, NP366–374/Db-specific T cells may be triggered by high levels of Ag in the lung airways to mediate effector functions and undergo rapid apoptosis. This would result in the observed skewing of the NP/PA ratio at the site of infection. Taken together, the poor viral clearance observed in PA224–233-vaccinated mice could not be explained by low numbers of Ag-specific T cells generated following infection or a reduced capacity to migrate to the site of infection.
Dendritic cell vaccination results in functionally active CD8+ T cells
The data thus far indicate that PA224–233-vaccinated mice generate greatly increased numbers of PA224–233/Db-specific T cells to influenza virus infection, but mediate delayed viral clearance. Thus, we investigated the functional properties of the NP366–374/Db- and PA224–233/Db-specific T cells generated under these conditions. As shown in Fig. 4, both NP366–374/Db- and PA224–233/Db-specific T cells produced IFN- and TNF- on day 10 postinfection as determined by an intracellular cytokine-staining assay. When calculated as absolute numbers of cells, the number of IFN--producing cells matched the numbers of cells determined by tetramer analysis. In contrast, only a subset of the IFN--producing cells produced TNF- in response to peptide stimulation. Interestingly, a much higher frequency of PA224–233/Db-specific T cells produced both IFN- and TNF- compared with NP366–374/Db-specific T cells (Fig. 4B) in response to the PA224–233 peptide than to the NP366–374 peptide (Fig. 4B). Whereas 30% of the NP366–374/Db-specific T cells produced TNF-, 75% of the PA224–233/Db-specific T cells produced TNF-. This is consistent with findings by others (19, 20, 21) that PA224–233/Db-specific T cells produce more TNF- upon restimulation than NP366–374/Db-specific T cells.
FIGURE 4. IFN- and TNF- production by Ag-specific CD8+ T cells on day 10 postinfection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31. On day 10 postinfection, the lungs were collected, enriched for lymphocytes, and incubated for 5 h in the presence of the indicated peptides and brefeldin A. Following stimulation, the cells were stained with anti-CD8 PerCP, anti-IFN- PE, and anti-TNF- allophycocyanin. A, The percentages of CD8 cells producing TNF- and IFN- following stimulation with either the NP366–374 or PA224–233 peptide. B, The number NP366–374/Db- and PA224–233/Db- specific cells as determined by tetramer staining, IFN- production, or TNF- production from one of three representative experiments.
In addition to determining the ability of T cells generated by vaccination to produce cytokines, we also determined their cytolytic activity. As seen in Fig. 5, following NP366–374 vaccination there were significantly higher numbers of NP366–374/Db-specific CTLs than PA224–233/Db-specific CTLs. Conversely, following PA224–233 vaccination there were significantly increased numbers of PA224–233/Db-specific CTLs compared with the number of NP366–374/Db-specific CTLs.
FIGURE 5. Ag-specific cytotoxicity on day 10 postinfection. C57BL/6 mice were administered 106 peptide-pulsed dendritic cells via i.v. injection, and on day 14 following vaccination, the mice were intranasally infected with 300 EID50 x31 influenza virus. On day 10 postinfection, the BAL was collected, enriched for lymphocytes, and incubated for 5 h with peptide-pulsed target cells labeled with 51Cr. Following incubation, the supernatants were collected and the percentage of specific release was determined. Shown is the percentage of specific release in NP366–374-vaccinated mice and PA224–233-vaccinated mice from one of three representative experiments.
Taken together, the data presented in this study show that vaccination with either NP366–374 or PA224–233 peptide-pulsed dendritic cells results in increased numbers of CD8+ T cells in the lungs following infection. These epitope-specific T cells were capable of producing IFN- and TNF- upon restimulation with Ag and had potent CTL activity. However, vaccination with the NP366–374 peptide resulted in accelerated viral clearance, whereas PA224–233 vaccination resulted in delayed viral clearance.
Discussion
We have investigated the impact of differential Ag presentation upon vaccine efficacy. As shown by others (9, 22, 23), we found that NP366–374 vaccination results in increased epitope-specific T cell numbers, increased peptide-specific IFN- and TNF- production, potent CTL activity, and accelerated viral clearance compared with unvaccinated mice. Conversely, we found that although PA224–233 vaccination resulted in increased numbers of epitope-specific T cells that are capable of migrating to the site of infection, producing IFN- and TNF- in response to Ag, and elaborating potent CTL activity, this response resulted in delayed viral clearance compared with that of unvaccinated mice. Together, these data support the idea that PA224–233/Db-specific T cells generated following vaccination are fully functional but may be inefficient at clearing virus due to poor expression of the PA224–233 epitope in the lungs. Importantly, the fact that PA vaccination actually resulted in delayed viral clearance compared with unvaccinated mice demonstrates that some epitopes may be detrimental in the context of a vaccine.
It is currently unclear why PA224–233-vaccinated mice exhibit a delay in viral clearance. Based on the data in Fig. 3, it appears likely that the enhanced PA224–233/Db-specific T cell response actually suppressed the NP366–374/Db-specific T cell response. Thus, PA224–233 vaccination may not only elicit T cells that do not mediate protection, but also decrease the production of protective NP366–374/Db-specific T cells (and probably protective T cells of other specificities). However, it should be noted that PA224–233-vaccinated mice were ultimately able to clear the viral infection (data not shown), presumably after the appearance of sufficient numbers of protective T cells specific for other influenza virus epitopes, such as NP366–374 and PB1703–711 (9, 22, 23). In this regard, the number of NP366–374/Db-specific T cells in the airways and lungs are roughly equivalent in unvaccinated and PA224–233-vaccinated mice by day 15 postinfection (data not shown).
One possible explanation for the reduced capacity of PA224–233/Db-specific T cells to clear virus is that they generate an inappropriate antiviral response. Interestingly, we found that although the majority (75%) of IFN--producing PA224–233/Db-specific T cells also produce TNF-, only 30% of the NP366–374/Db-specific T cells produce both IFN- and TNF-. Recent reports by Turner et al. (21) and La Gruta et al. (20) show that following infection of normal C57BL/6 mice, the PA224–233/Db-specific T cell population makes more IFN- and TNF- following short-term in vitro restimulation than the NP366–374/Db-specific T cell population due to higher functional avidity of Ag binding by PA224–233/Db-specific TCR. This difference in TNF- production results in a selective TNFR2-mediated deletion of high avidity PA224–233/Db-specific T cells following in vitro restimulation with high doses of Ag (21). However, this deletion is less apparent in the T cell populations recovered from the lungs of influenza virus-infected mice (21). In this regard, we find that rather than TNF--mediated deletion of PA224–233/Db-specific T cells following infection of vaccinated mice, there is a significant increase in the number of PA224–233/Db-specific in the lungs. Additionally, although this vaccination strategy results in large numbers of the high avidity PA224–233/Db-specific T cells, there is delayed viral clearance. This correlates with our previous data that the lung epithelial cells poorly present the PA224–233/Db epitope, which would result in low Ag levels following infection, survival of the high affinity PA224–233/Db-specific T cells at this site, and delayed viral clearance.
It is currently unclear whether the differential presentation of the PA224–233/Db epitope by dendritic and nondendritic cells is a unique situation or representative of a class of epitopes. Recently, Zhong et al. (24) identified an array of previously unidentified CD8+ T cell epitopes in the PR8/8/34 strain of influenza virus and examined the ability of epitope-specific CD8 T cells to lyse peptide-pulsed or virally infected EL4 target cells. Interestingly, the study identified two new epitopes that, like the PA224–233 epitope, do not appear to be presented by virally infected EL4 cells. Thus, there is an indication that differential Ag processing and presentation may occur with other influenza virus epitopes.
Taken together, the data in this report show that dendritic cell vaccination resulted in increased numbers of epitope-specific CTLs capable of migrating to the site of infection and producing IFN- and TNF-. Additionally, we have shown that vaccination with an epitope presented predominantly by infected dendritic cells, but not lung epithelial cells, resulted in delayed control of a subsequent viral infection. Understanding the impact of differential Ag processing is therefore essential for the development of effective peptide-based vaccines.(Sherry R. Crowe, Shannon )