|Home | About | Journals | Submit | Contact Us | Français|
Vaccination with formalin-inactivated respiratory syncytial virus (RSV) vaccine results in enhanced respiratory tract inflammation and injury following subsequent RSV infection. RSV vaccine-enhanced disease can also be produced in mice by prior vaccination with a vaccinia virus vector containing the RSV G protein, followed by intranasal infectious RSV challenge, a process characterized by induction of a potent memory CD4+ T-cell response to challenge infection with some features characteristic of Th-2 CD4+ T-cell responses, including increased eosinophil accumulation in pulmonary inflammatory infiltrates. The adaptive immune response to the RSV G protein in immunized BALB/c mice is characterized by a weak or absent primary and secondary recall CD8+ T-cell response. These and related results have led to the hypothesis that the failure of the infected animals to mount an effective CD8+ memory T-cell (CD8+ Tm) response in this model could account for the pulmonary eosinophilia associated with the development of enhanced disease, and that CD8+ T cells may control the development of eosinophilia. In this study, we investigated how and when the generation of a CD8+ Tm response to RSV infection might affect the development of pulmonary eosinophilia in this model of vaccine-enhanced disease. By defining the CD8+ T-cell response kinetics and monitoring lung parenchymal eosinophil accumulation, we show that the establishment of an RSV-specific CD8+ Tm response in the infected lungs early after challenge infection (i.e., within the first 3d of RSV infection) is necessary and sufficient to control pulmonary eosinophilia development. Additionally, our work suggests that the mechanism by which CD8+ T cells regulate this process is not by modulating the differentiation or development of the CD4+ Tm response. Rather, we demonstrate that IL-10 produced by early responding CD8+ Tm cells may regulate the pulmonary eosinophilia development observed in RSV vaccine-enhanced disease.
Respiratory syncytial virus (RSV) is a major cause of infant morbidity from severe lower respiratory tract infection with bronchiolitis worldwide. Early attempts at vaccination of seronegative infants and young children using formalin-inactivated RSV were unsuccessful. Vaccinees, after subsequent natural RSV infection, exhibited a more severe illness characterized in several instances by enhanced pulmonary injury with extensive mononuclear cell infiltration and pulmonary eosinophilia, compared to placebo vaccine control children (4,8,9). Vaccination of BALB/c mice with the formalin-inactivated RSV vaccine or RSV proteins (including RSV G protein) expressed from a recombinant vaccinia virus also results in enhanced pulmonary injury and eosinophilia after intranasal RSV challenge infection, although the exact mechanisms for eosinophilia and enhanced pulmonary diseases in the two models may actually differ (7,11,15,23). More recent studies demonstrated that the majority of CD4+ T cells responding during RSV infection in RSV G-primed BALB/c mice recognize a single MHC class II (I-Ed) restricted immunodominant epitope within the RSV G protein spanning amino acid residues 185–193, and also predominantly express the Vβ14 T-cell receptor (21,22). Additionally, Vβ14+ CD4+ T cells were shown to be the critical population responsible for inducing the pulmonary eosinophilia observed in RSV-infected RSV G-primed BALB/c mice (21).
Interestingly, in this murine model, immunization with RSV G protein results in a negligible to absent CD8+ memory T-cell (CD8+ Tm) response to secondary challenge RSV infection. It has thus been hypothesized that the lack of a CD8+ Tm response to challenge RSV infection in BALB/c mice vaccinated with formalin-inactivated RSV vaccine or RSV G protein is linked to the development of pulmonary eosinophilia associated with RSV vaccine-enhanced disease. In support of this theory, it has been demonstrated that, in RSV G-primed mice, when a CD4+ Tm and CD8+ Tm response is simultaneously induced in response to challenge RSV infection, pulmonary eosinophilia is markedly diminished (14,17).
In this report, we investigated how and when the generation of a CD8+ Tm response to RSV infection may affect the development of pulmonary eosinophilia in this model of RSV vaccine-enhanced disease. By defining the kinetics of the CD8+ T-cell response, as well as monitoring the accumulation of eosinophils in the lung parenchyma, we show that the development of the RSV-specific CD8+ Tm response in the infected lungs early after challenge infection (i.e., within the first 3d of RSV infection) is necessary and sufficient to control the development of pulmonary eosinophilia. Additionally, our work suggests that the mechanism by which CD8+ T cells regulate this process is not by modulating the differentiation and/or development of the CD4+ Tm response. Rather, we demonstrate that IL-10 produced by early-responding CD8+ Tm cells may regulate the development of pulmonary eosinophilia observed in RSV vaccine-enhanced disease.
Female BALB/c AnNTac (H-2d) mice were purchased from Taconic Farms (Germantown, NY) and immunized (as described below) at 8–10wk of age for all experiments. Male IFN-γ−/− BALB/c mice were purchased from the Jackson Laboratory (Bar Harbor, ME). The mice were housed in a pathogen-free environment and all animal experiments were performed in accordance with protocols approved by the University of Virginia Animal Care and Use Committee.
Recombinant vaccinia viruses expressing either the RSV G protein (vvG) or the RSV M2 protein (vvM2) were gifts from J.L. Beeler (Food and Drug Administration, National Institutes of Health, Bethesda, MD). Recombinant vaccinia viruses expressing β-galactosidase (vvβgal) were used as controls for immunization. The A2 strain of RSV (a gift of P.L. Collins, National Institute of Allergy and Infectious Disease, National Institutes of Health) was grown in HEp-2cells (American Type Culture Collection, Manassas, VA) and titered for infectivity. The preparation of influenza virus A/PR/8/34 (H1N1) was described elsewhere (25). The RSV viral peptide M282–90 was synthesized by the University of Virginia Biomolecular Research Facility.
Groups of four mice were vaccinated with 3×106 PFU vvG, vvM2, or vvβgal, in a 10-μL volume via scarification at the base of the tail with a 25-gauge needle. After a minimum of 7wk, these mice were lightly anesthetized with halothane (Halocarbon Laboratories, Rivers Edge, NJ) and then intranasally challenged with 2.16×106 PFU RSV in a 50-μL volume or with 0.1 LD50 A/PR/8/34 in a 50-μL volume. At various times post-infection (p.i.), mice were sacrificed and bronchoalveolar lavage fluid, peri-bronchical lymph nodes (PBLN), lung, and spleen samples were collected for further analyses.
Anti-IL-10Rα (1B1.3a) mAbs or rat IgG1 control mAbs were obtained from Bio-Express, West Lebanon, NH. Briefly, 750μg of anti-IL-10Rα mAbs or Rat IgG1 control mAbs in 500μL 1× sterile PBS was injected intraperitoneally (IP) into vaccinated BALB/c mice at day −1 and day +1 of RSV infection.
The CD8α depletion antibody was isolated and purified from the 2.43 hybridoma cell line by the University of Virginia Lymphocyte Culture Center; then 100μg of anti-CD8α mAbs or of functional grade-purified mouse IgG2b isotype control (eBioscience, San Diego, CA) in 500μL 1× sterile PBS was injected IP into vaccinated BALB/c mice at different days prior to or following RSV infection as indicated.
CD8+ T cells were purified from the spleens of various vaccinated and naïve mice using the AutoMACS Separator (Miltenyi Biotec GmbH, Auburn, CA) according to the manufacturer's protocol. CD8+ T cells were positively selected using anti-CD8α microbeads (Ly-2; Miltenyi Biotec GmbH) with a greater than 85% purity confirmed by flow cytometry. Approximately 1×106 purified CD8+ splenocytes were injected IV into vvG-primed BALB/c hosts at various times prior to or following RSV infection as indicated.
Lungs were flushed via the right ventricle with approximately 4mL of 1× PBS with 10U/mL of heparin (Sigma, St. Louis, MO) to remove blood. The lungs were then removed aseptically and separated away from the heart, thymus, and bronchial lymph nodes. The lungs were minced in RPMI media (GIBCO BRL, Gaithersburg, MD) supplemented with 10U/mL penicillin (Gibco BRL) and 10μg/mL streptomycin (Gibco BRL) and passed through a wire screen followed by a quick centrifugation at 400 rcf to remove particulates Lung mononuclear cell suspensions were counted using a hemacytometer.
The spleens were aseptically removed, placed in complete medium, and passed through a wire screen. Following a quick centrifugation at 400 rcf to remove particulates, spleen mononuclear cell suspensions were counted using a hemacytometer.
Mediastinal and peri-bronchical lymph nodes were dissected away from the lungs and placed in complete medium. The nodes were gently disrupted between the frosted ends of two glass microscope slides and the slides were then extensively rinsed to optimize cell recovery. Mononuclear cell suspensions were counted using a hemacytometer.
The mice were sacrificed and an incision was made in the skin over the larynx. A cannula attached to a syringe was inserted at the incision site and 500μL of complete media was flushed through the lungs three times. The BAL mononuclear cell suspensions recovered were spun at 1400rpm for 5min. Afterwards, supernatants were collected and stored at −80°C until use.
Supernatants of re-stimulated PBLN and lung cultures in addition to supernatants of directly isolated BAL fluid were screened by either BD OptEIA™ mouse IFN-γ (AN-18) ELISA or BD OptEIA™ mouse IL-10 ELISA (BD Biosciences Pharmingen) according to the manufacturer's protocol. Briefly, 96-well flat-bottom immuno-plates (Nalge Nunc International, Rochester, NY) were coated with anti-cytokine coating mAbs overnight at 4°C. The next day, the plates were washed with 1× PBS-0.05% Tween-20 (Sigma) and blocked with PBS-10% FCS for a minimum of 2h. After blocking, the plates were washed again and 50μL/well of sample or recombinant cytokine standard was added in duplicate and incubated at room temperature for 2h. The plates were then washed and incubated for 1h at room temperature in the presence of biotinylated anti-cytokine mAbs. The ELISA plates were developed using 1× TMB substrate solution (eBioscience) and the reaction was stopped with 1M H2SO4. Plates were read at 450nm using a Powerwave XS plate reader and analyzed using KC Junior software (both from Bio-Tek Instrument Inc., Winooski, VT).
For multicolor FACS analysis, approximately 2×106 cells from harvested PBLN, spleen, or lung suspensions were incubated for 5min with Fc receptor (2.4G2) blocking agent in 100μL of FACS buffer (1× PBS supplemented with 2% heat-inactivated fetal calf serum [Atlanta Biologicals, Norcross GA] and 0.02% NaN3) to prevent nonspecific binding. Next, specific mAbs were directly added to the samples and incubated for 45min in the dark at 4°C. Antibodies used were at a final concentration of 1μg/100μL and included PE-conjugated anti-Siglec-F (E50-2440), PerCP-Cy5.5 conjugated CD45 (30-F11), FITC-conjugated CD11c (HL3), APC-conjugated CD11b (M1/70), FITC-conjugated Vb14 TCR (14-2), PerCP-Cy5.5 conjugated CD8α (53-6.7), APC-conjugated CD4 (L3T4), PE-conjugated IFN-γ, and APC-conjugated IL-10 (the antibodies were purchased from BD Pharmingen). The PE-conjugated H-2Kd RSV M2 tetramer was purchased from the Baylor tetramer facility (Baylor College of Medicine, Houston, TX). Stained cells were fixed and lysed with FACS lysing solution (Becton Dickinson, San Jose, CA), then washed and resuspended in FACS buffer. Stained cell suspensions were analyzed using either a FACScalibur or a FACScanto flow cytometer (BD Biosciences) and FlowJo software (TreeStar, Ashland, OR).
All data are expressed as the mean±standard error of the mean. When differences between groups were assessed, an ANOVA supported by a Tukey's multiple comparison test was used with the null hypothesis being rejected when p<0.05.
As noted above, the development of pulmonary eosinophilia in RSV G-primed mice undergoing challenge RSV infection is dependent upon the development of a memory CD4+ T-cell response to the G protein (21). We have recently reported that the orchestration of an early RSV G-specific CD4+ Tm response in the lungs to challenge RSV infection may be important in the development of the pulmonary eosinophilia associated with the enhanced lung injury (24). Furthermore, the simultaneous induction of a recall (memory) CD8+ T-cell (CD8+ Tm) response in this model of challenge RSV infection attenuates the development of pulmonary eosinophilia (17). However, the mechanism(s) of CD8+ Tm-mediated control of eosinophil production in the lungs has not been fully elucidated. In view of this evidence suggesting a role for the early memory G-specific CD4+ T-cell response in the development of pulmonary eosinophilia (24), we analyzed the kinetics of the early CD8+ Tm response in the lungs and draining PBLN of immune mice undergoing challenge RSV infection. To evaluate the tempo of the induction of the CD8+ Tm response early in challenge infection, we vaccinated BALB/c mice with a vaccinia virus expressing the RSV M2-1 protein (vvM2). The M2-1 protein contains the immunodominant M282–90 peptide epitope recognized by CD8+ T cells in this mouse strain (2,17). We enumerated (by MHC class I tetramer staining) RSV M2-specific CD8+ T cells in the lungs and PBLN of M2-immune mice at various time points following challenge RSV infection. The controls for this analysis included (1) RSV M2-immune mice infected with type A influenza, and (2) mice vaccinated with a vaccinia virus vector expressing β-galactosidase that subsequently underwent a primary RSV infection.
As Fig. 1A demonstrates, we could detect M2-tet+ CD8+ T cells in the lungs of vvM2-primed mice as early as 1d post challenge RSV infection. M2-tet+ CD8+ T-cell numbers gradually increased between day 1 and day 3p.i. (from ~5×104 M2-tet+ cells at day 1, up to ~2.6×105 cells at day 3). This was followed by a 10- to 30-fold increase in RSV-specific (M2-tet+) CD8+ T cells by days 4–5p.i. (Fig. 1A). This early (days 1–3p.i.) recruitment of RSV-specific CD8+ Tm was not observed when vvM2-immune mice were intranasally challenged with type A influenza virus (data not shown). Likewise, control mice primed with a vaccinia virus expressing β-galactosidase (vvβ-gal) undergoing a primary intranasal RSV infection demonstrated no detectable increase in M2-tet+ CD8+ T cells in the infected lungs until day 5p.i. (Fig. 1A). Thus, the recruitment or expansion of this population of CD8+ Tm in the lungs was antigen-specific and not affected by a non-specific inflammatory stimulus to lungs (e.g., infection with an unrelated virus).
In a companion analysis of the CD8+ Tm response in the draining PBLN, M2-tet+ CD8+ T cells were similarly detected as early as day 1p.i. (Fig. 1B). The numbers of M2-tet+ CD8+ T cells rapidly increased between days 2 and 3p.i., and cell numbers subsequently declined at day 4p.i. (Fig. 1B). This fall in the number of RSV-specific CD8+ T cells in the draining PBLN coincided with the rapid increase in the number of M2-tet+ CD8+ T cells in the infected lungs (i.e., between days 3 and 4p.i.) (Fig. 1A), and suggested that the drop in lymph node–resident RSV-specific CD8+ Tm cells was likely due to the recruitment of the proliferating T cells from the draining nodes into the infected lungs.
Effector and memory CD8+ and CD4+ T cells produce a variety of inflammatory mediators in response to antigen recognition at sites of infection. The pro-inflammatory cytokine IFN-γ is an important mediator produced by CD8+ T cells in response to antigen. To assess the relationship between the tempo of CD8+ Tm accumulation in the infected lungs and the production of pro-inflammatory cytokines, we examined the early time course of IFN-γ release into the BAL fluid of vvM2-immune mice undergoing challenge RSV infection. As Fig. 2 demonstrates, we could detect significant IFN-γ production by day 2p.i., which increased three- to fourfold by day 3p.i., a result in accordance with the expanded number of virus-specific (M2-tet+) CD8+ Tm cells accumulating in the infected lungs. We could also detect low levels of IFN-γ release into the BAL fluid as early as day 1p.i. in the M2-immune RSV-challenged mice (Fig. 2). Elimination of CD8+ T cells prior to infection by in-vivo administration of a depleting anti-CD8α antibody inhibited the release of IFN-γ into the BAL fluid, suggesting that CD8+ T cells were the primary source of the IFN-γ detected at this early time during infection (Fig. 2). As expected, the bulk of the IFN-γ produced was dependent on specific viral antigen recognition, since vvM2-immune mice infected with influenza virus released low levels of IFN-γ over the same time frame (Fig. 2).
Mice primed with vvG mount a strong recall CD4+ Tm response to challenge RSV infection with minimal or no detectable CD8+ Tm response (14,17). In order to provide a source of both CD4+ Tm and CD8+ Tm for analysis of the regulation of pulmonary eosinophilia, we simultaneously vaccinated mice with vvG and vvM2. When these dual vaccinated mice underwent challenge RSV infection, the kinetics of the recall M2-specific CD8+ Tm response in the lungs and PBLN was comparable to the response of mice undergoing challenge RSV infection after vaccination with vvM2 alone (Fig. 1A).
We next examined the impact of CD8+ Tm depletion on the development of pulmonary eosinophilia. Eosinophils in the lungs of vvG/vvM2-primed mice undergoing challenge RSV infection were quantitated by flow cytometric analysis of eosinophil numbers in total infected lung cell suspensions as previously described (18). As previously observed (6,17), elimination of CD8+ T cells in vvG/vvM2-primed mice 1d prior to challenge RSV infection by in-vivo administration of a depleting anti-CD8α mAb resulted in the increased accumulation of eosinophils when total lung cell suspensions were analyzed at days 3 and 5p.i. (Fig. 3A). Eosinophil accumulation was well above the cell numbers observed in RSV-infected immune mice who either received an isotype control antibody or no treatment (Fig. 3A). Additionally, an increase in lung eosinophil numbers was observed as early as day 2 post-challenge infection in CD8+ T-cell depleted mice (data not shown).
It is noteworthy that in the untreated, depleting antibody-treated, and control antibody-treated RSV-infected recipients, lung eosinophil numbers were maximum by day 3p.i. This finding raised the possibility that maximum eosinophil recruitment occurred early in the recall response, and that the CD8+ Tm recruited to the lungs early in response to challenge infection may play a critical role in regulating (suppressing) the development of pulmonary eosinophilia. To explore this, we examined how the timing of CD8+ T-cell depletion relative to challenge RSV infection impacted eosinophil accumulation. We found that the elimination of the CD8+ T cells immediately prior to infection (Fig. 3B) resulted in the expected increased eosinophil accumulation in the infected lungs. However, depletion of CD8+ T cells between days 3 and 4 did not result in increased pulmonary eosinophilia (Fig. 3B), in spite of the fact that RSV-specific CD8+ Tm were rapidly increasing in numbers in the infected lungs at this time (Fig. 1A).
The above findings suggested that the CD8+ T cells responding to infection early in the recall response (i.e., between days 1 and 3) were responsible for suppressing the development of pulmonary eosinophilia. We had previously reported in this model that CD4+ Tm directed to an immuno-dominant epitope on the RSV G protein, which employed a TCR complex utilizing the Vβ14 TCR gene, orchestrated the development of pulmonary eosinophilia in response to challenge infection (21). When we examined the impact CD8+ T-cell depletion had on the accumulation of CD4+ Vβ14+ T cells in the infected lungs, there was no preferential increase in the accumulation of these CD4+ Tm in the CD8+ T-cell depleted recipients at day 3p.i. (Fig. 3C). There was a statistically significant increase in the total number of these CD4+ Vβ14-expressing T cells at day 5p.i. in infected anti-CD8α-treated mice. However, this increase in RSV G-specific memory CD4+ T cells occurred after the maximum level of pulmonary eosinophilia was observed in the CD8+ T cell–depleted animals.
IFN-γ is well recognized as an important regulator of CD4+ T-cell differentiation into Th-1-type effector T cells, and would represent a potential modulator of pulmonary eosinophilia in challenge RSV infection through the effect of this cytokine on CD4+ Tm. However, the contribution of IFN-γ (and in particular T cell–derived IFN-γ) in controlling eosinophil accumulation in the lungs of vvG-primed mice undergoing challenge RSV infection is not certain, with the most recent evidence suggesting a minimal role for this cytokine in regulating pulmonary eosinophilia in this model (14). Since we could detect substantial production of IFN-γ by RSV-specific CD8+ Tm as early as day 2 following challenge RSV infection in RSV-M2-immune mice (Fig. 2), and since the above findings suggested an important role for CD8+ Tm responding early in infection in controlling lung eosinophilia, we evaluated the contribution of CD8+ Tm-derived IFN-γ in regulating pulmonary eosinophilia.
To do so, we adoptively transferred purified CD8+ T cells from RSV-M2-immune wild-type or IFN-γ-deficient donors into eosinophil prone RSV G-immune mice. One day after CD8+ T-cell transfer, the three groups of RSV G-immune mice [i.e. CD8+ IFN-γ deficient (IFN-γ−/−) T cell recipients, CD8+ IFN-γ sufficient (IFN-γ+/+) T cell recipients, and untreated immune control mile] underwent challenge RSV infection. As Fig. 4 demonstrates, both transferred wild-type CD8+ T cells, as well as CD8+ Tm from IFN-γ-deficient donors, were equally capable of suppressing the development of pulmonary eosinophilia at day 5 post-challenge infection in the RSV G-immune animals. These results support the view that CD8+ T-cell-derived IFN-γ does not play an essential role in regulating the development of eosinophilia in this model of RSV vaccine-enhanced disease.
IL-10 is a regulatory cytokine that can act on a variety of cell types to suppress pro- inflammatory responses and limit the extent of inflammation seen during infection and other types of tissue injury (5,12). IL-10 is generally considered to be a product of distinct subsets of innate immune effector cells (e.g., regulatory DCs, NK cells, and NK T cells), and of regulatory and effector CD4+ T cells (3). We have, however, recently demonstrated that IL-10 is produced by activated effector CD8+ T cells responding in the lung to type A influenza virus in the murine model of acute intranasal influenza virus infection (19). Since IL-10 has been reported to regulate eosinophil function and accumulation in allergic responses (5), we asked whether IL-10 is produced by RSV-specific CD8+ Tm early in the recall response of vvG/vvM2-primed mice to challenge RSV infection.
As Fig. 5A demonstrates, IL-10 was detected in the BAL fluid of vvG/vvM2-primed mice at days 2 and 3p.i. Furthermore, in-vivo depletion of CD8+ T cells in these mice by monoclonal antibody treatment prior to challenge infection suppressed the production of IL-10. In keeping with this finding, we were able to demonstrate the presence of IL-10-secreting RSV-M2-specific CD8+ T cells in the lungs of vvG/vvM2-primed mice undergoing challenge RSV infection at day 3p.i. (Fig. 5B), after short-term stimulation of lung cells with M282–90 peptide in the in-vitro intracellular cytokine assay. As we recently observed for influenza infection (19), the IL-10-producing lung CD8+ T-cell effectors represent a subset of CD8+ T cells that simultaneously produced both IL-10 and IFN-γ.
To determine if CD8+ T-cell-derived IL-10 had an impact on the accumulation of eosinophils in the vvG/vvM2-immune mice undergoing challenge infection, we examined the effect of blocking IL-10 action in vivo by administration of a blocking monoclonal anti-IL-10R antibody at the time of challenge infection. As Fig. 5C demonstrates, there was a significant increase in the number of eosinophils detected in the lung parenchyma (as well as in the BAL fluid; data not shown) in the infected mice treated with the blocking anti-IL-10-receptor antibody over that observed in infected animals receiving an isotype control antibody.
In this report, we examined the role of RSV-specific CD8+ Tm in regulating the development of pulmonary eosinophilia in the recall response to challenge RSV infection in RSV G-immune mice. We found that CD8+ Tm directed to the RSV-M2 protein start to accumulate in the infected lungs early after challenge infection—that is, between days 1 and 3p.i. Furthermore, by eliminating CD8+ T cells in vivo before or during challenge infection, we were able to demonstrate that the CD8+ Tm accumulating in the lungs during this early phase of infection play a critical role in regulating eosinophil accumulation in the infected lungs. In addition, we provide evidence that one mechanism by which CD8+ Tm may control the development of pulmonary eosinophilia in this model is through CD8+ Tm production of the regulatory cytokine IL-10.
We could detect the presence of RSV M2-specific CD8+ Tm in the lungs of immune mice undergoing challenge RSV infection as early as day 1p.i. by tetramer staining. The numbers of CD8+ Tm increased progressively up to day 3p.i. in the infected lungs. As expected, this early increase in virus-specific CD8+ T-cell numbers in the lungs was not observed in vvβ-gal primed mice undergoing primary RSV infection. There was a parallel expansion in the number of M2-tet+ CD8+ T cells in the PBLN of the immune mice up to day 3p.i. One attractive explanation for the presence of these virus-specific CD8+ Tm in the lungs early in infection is that they were derived from a population of circulating CD8+ Tm, which entered the draining PBLN, activated there, and rapidly exited the draining lymph nodes to the infected lungs, where they subsequently underwent additional proliferative expansion. We recently reported that such a mechanism might account for the early accumulation of CD4+ Tm in the infected lungs during challenge RSV infection (24). In this connection, a significant percentage (up to 15–30%) of the RSV-M2-specific CD8+ Tm found in the infected lungs at days 1–3 post-challenge infection contained >2N DNA content (data not shown), suggesting that these infiltrating memory CD8+ T cells were actively dividing. The RSV-specific CD8+ T cells present in the infected lungs early in the recall response could be also derived from circulating Tm that entered the infected lungs directly from the circulation and underwent proliferative expansion there. Our findings do not formally distinguish between these possibilities. However, we do believe that the rather extensive expansion of CD8+ Tm in the lungs between days 3 and 5p.i. (in part, at least) reflects the mobilization of activated memory effector CD8+ T cells from the draining PBLN into the infected lungs. The corresponding apparent decline in the numbers of virus-specific CD8+ T cells in the draining PBLN between days 3 and 4p.i. is most easily explained by such a mechanism.
We observed that, along with the early rapid accumulation and recruitment of CD8+ Tm cells into the infected lungs of immune mice, there was a corresponding release in the pro-inflammatory cytokine IFN-γ (as detected in BAL fluid). In vvM2-immune mice undergoing challenge infection, in-vivo depletion of CD8+ cells by administration of a depleting antibody indicated that CD8+ T cells were the primary source of this cytokine. IFN-γ release into the BAL fluid increased progressively up to day 4 post-challenge infection, and this increase in cytokine production directly paralleled the increase in CD8+ Tm accumulation in the infected lungs over the same time period. Indeed, there was IFN-γ release into the BAL fluid (above the background level observed in RSV-M2-immune mice undergoing challenge influenza infection) as early as day 1 post-RSV challenge, suggesting that sufficient numbers of CD8+ Tm were present in the infected lungs and were capable of rapidly responding, even at this early time point.
In this model of vaccine-enhanced disease, pulmonary eosinophilia as determined by eosinophil accumulation in the BAL fluid reaches a maximum at days 5–6 post-challenge RSV infection (22). However, in this report, we observed that the antecedent accumulation of eosinophils into the lung interstitium reaches a maximum at day 3p.i. Therefore, the factors which regulate the development of pulmonary eosinophilia in this model such as the Vβ14+ RSV G-specific CD4+ Tm (22), and as demonstrated here and elsewhere (14) the RSV-specific CD8+ Tm cells, likely operate early in the recall response to regulate eosinophil accumulation in the RSV-infected lungs. Our recent report (24), demonstrating the recruitment of Vβ14+ RSV G-specific CD4+ Tm into the infected lungs early in the recall response, as well as corresponding findings for memory CD8+ Tm reported here, are consistent with this concept.
Additional support for the critical role of the early CD8+ T-cell response in regulating the extent of pulmonary eosinophilia came from our analysis of the impact of the elimination of CD8+ T cells in vvG/vvM2-primed mice undergoing challenge RSV infection. In-vivo depletion of CD8+ T cells just prior to challenge infection results (as previously reported in 6,14,17) in the increased accumulation of eosinophils in the infected lungs. By contrast, depletion of CD8+ T cells between days 3 and 4p.i. does not reverse the inhibitory effect of the recall CD8+ T-cell response on the development of pulmonary eosinophilia, even though the majority of the RSV-specific CD8+ Tm cells accumulate in the lungs between days 3 and 5p.i.
The mechanisms through which CD8+ T cells regulate the development of pulmonary eosinophilia in this model of RSV vaccine-enhanced disease are not completely understood. Because of its well-documented effect on the differentiation of CD4+ T cells into Th-1 effectors, IFN-γ produced by CD8+ Tm cells early during challenge infection (i.e., days 1–3p.i.) would represent an attractive candidate molecule to suppress pulmonary eosinophilia through its effect on CD4+ Tm differentiation. However the findings reported here and elsewhere (14) argue against a critical role for T cell-derived IFN-γ in mediating the suppressive effect of responding CD8+ T cells on the development of pulmonary eosinophilia.
We recently reported (19) that during experimental murine influenza infection, high levels of the regulatory cytokine IL-10 are produced by responding effector CD8+ T cells in the infected lungs. The kinetics of IL-10 release into the infected respiratory tract is identical to that of IFN-γ, and the source of the IL-10 is a subset of effector T cells that simultaneously produce both IFN-γ and IL-10. Because of the well-documented effect of IL-10 on the recruitment and activity of inflammatory cells including eosinophils, we analyzed IL-10 release during the early phase (days 2–3) of the recall response to challenge infection of vvG/vvM2-primed mice. We found that IL-10 is released into the respiratory tract of vaccinated mice in the early phase of challenge RSV infection, and that the IL-10 release is dependent on CD8+ cells. As observed in the primary influenza infection (19), the kinetics of IL-10 released in the airway is similar to that of IFN-γ, suggesting that the IL-10 most likely comes directly from the CD8+ T cells themselves. However, based on the available evidence, we cannot exclude the possibility that the IL-10 detected is produced by another cell type within the RSV-infected respiratory tract, whose production of IL-10 requires interaction with responding RSV-specific memory CD8+ T cells. We also note that the percentage of IL-10-producing CD8+ T cells in the antigen-specific CD8+ T-cell population in this model is relatively low compared to that of IL-10-producing CD8+ T cells seen during acute influenza infection (19). The exact mechanism of this discrepancy remains unknown, but it may reflect the difference in overall inflammatory stimuli produced in the lung between the two viral infections. Nevertheless, we found that blocking the action of this T-cell-derived IL-10 in vivo by administration of a blocking anti-IL-10 receptor antibody results in increased pulmonary eosinophilia, similar to the effect of in-vivo CD8+ T-cell depletion.
The release of IL-10 by RSV-specific effector CD8+ Tm during challenge RSV infection represents an attractive mechanism to account for the inhibitory effect of the CD8+ Tm response on pulmonary eosinophilia in this model. IL-10 can suppress the production and/or action of a variety of cytokines and chemokines and expression of adhesion molecules, including mediators and molecules implicated as critical regulators of eosinophil recruitment and extravasation into the lung interstitium (1,10,16,20). In this connection, however, we would emphasize that the CD8+ T cell-derived IL-10 is unlikely to be acting selectively on eosinophil recruitment or accumulation. Rather, the increase in lung eosinophil accumulation with in-vivo IL-10 receptor blockade more likely reflects an increase in the extent of pulmonary inflammation (including increased eosinophil accumulation in the infected lungs) when the anti-inflammatory effect of IL-10 is blocked. Furthermore, our data may also provide an explanation that the number of CD8+ T cells in the lung determines the ability of CD8+ T cells to inhibit RSV vaccine-enhanced pulmonary eosinophilia (13). In this case, we think that the number of CD8+ T cells specific for the subdominant epitope within the RSV fusion (F) protein is sufficient to produce a large amount of IL-10 to inhibit the early recruitment of eosinophils after RSV challenge infection.
In conclusion, in this report we have examined the role and effect of the memory CD8+ T cells on the development of pulmonary eosinophilia in a murine model of vaccination-induced RSV-mediated enhanced pulmonary injury. Our results suggest that CD8+ Tm recruited to the infected lungs early in the recall response to challenge infection play an important regulatory role in suppressing the development of pulmonary eosinophilia. Furthermore, we provide evidence that one potential mechanism of control of eosinophil accumulation by the responding CD8+ T cells is the production of IL-10 by the responding CD8+ T cells, which in turn acts to control the level of lung inflammation.
This work was supported by the National Institutes of Health (grants AI15608, HL-33391, and AI37293 to T.J.B.). The authors would like to thank all the members of the Braciale lab for their assistance.
The authors declare that they have no conflicting financial interests.