Eosinophils do not contribute to respiratory syncytial virus vaccine-enhanced disease

Abstract

Respiratory syncytial virus (RSV) infection of BALB/c mice previously immunized with a recombinant vaccinia virus (vacv) expressing the attachment (G) protein of RSV results in pulmonary eosinophilia, which mimics the response of FI-RSV vaccinated children, as well as increased weight loss, clinical illness, and Penh. We show that RSV infection of eosinophil-deficient mice previously immunized with vacvG results in the development of increased weight loss, clinical illness, and Penh that is similar to wild-type controls. These measures of RSV vaccine-enhanced disease are dependent upon STAT4. Interestingly, neither IL-12 nor IL-23, the two most common STAT4-activating cytokines, proved necessary for the development of disease. We demonstrate that IFN-γ, which is produced following STAT4 activation, contributes to clinical illness and increased Penh, but not weight loss. Our results have important implications for future RSV vaccine design suggesting that enhancing a Th1 response may exacerbate disease.

Introduction

Respiratory syncytial virus (RSV)³ infection is the leading cause of lower respiratory tract disease in infants and young children (1). Children administered a formalin-inactivated (FI) RSV vaccine experienced increased morbidity and mortality upon subsequent natural RSV infection as compared to children that received a control FI-parainfluenza virus vaccine (2, 3). Histological examination at autopsy revealed extensive pulmonary inflammation with the presence of numerous eosinophils (2). In addition, elevated numbers of eosinophils were also noted in the peripheral blood of approximately 55% of the vaccines examined (3). Based on the previous experience with FI-RSV immunization in humans, the development of pulmonary eosinophilia has become the hallmark of RSV vaccine-enhanced disease. However, the extent to which eosinophils directly contribute to the pathogenesis of RSV vaccine-enhanced disease has not been firmly established.

We have previously demonstrated that the Th2 cell-associated cytokine IL-13 is necessary for the development of pulmonary eosinophilia upon RSV challenge of mice previously immunized with a recombinant vaccinia virus (vacv) expressing the attachment (G) protein of RSV (vacvG) (4). In this report, we examined STAT6-deficient mice to determine if either IL-4 or IL-13 contributed to the development of increased weight loss, clinical illness, and Penh after RSV challenge of mice previously immunized with vacvG. We show that STAT6 is necessary for the development of pulmonary eosinophilia, but that it is not required for increased weight loss, clinical illness, and Penh after RSV challenge of mice previously immunized with vacvG. Because eosinophilia is considered the hallmark of RSV vaccine-enhanced disease, we verified our results using ΔdblGATA mice, which are deficient in the generation of eosinophils (5). At the same time, we examined STAT4-deficient mice to determine the contribution of Th1 cell-associated signaling to the development of RSV vaccine-enhanced weight loss, clinical illness, and Penh. Our results demonstrate that the transcription factor STAT4 is required for the development of weight loss, clinical illness, and increased Penh. Assessing the role of IFN-γ, a cytokine produced in response to STAT4 activation, our data indicates that the production of IFN-γ contributes to clinical illness and increased Penh, but not weight loss. Our results demonstrate that eosinophils are not required for the development of increased weight loss, clinical illness, or Penh after RSV challenge of mice previously immunized with vacvG.

Materials and Methods

Mice

BALB/cAnNCr mice were purchased from the National Cancer Institute. STAT4-deficient mice (C.129S2-Stat4^tm1Gru/J), STAT6-deficient mice (C.129S2-Stat6^tm1Gru/J), IL-12p40-deficient mice (C.129S1-Il12b^tm1Jm/J), IFN-γ-deficient mice (C.129S7(B6)-Ifng^tm1Ts/J), and ΔdblGATA mice (C.Cg-Gata1^tm6Sho/J), all on a BALB/c background, were purchased from The Jackson Laboratory. All experimental procedures were approved by the University of Iowa Animal Care and Use Committee.

Viruses

Recombinant vacv and the A2 strain of RSV were propagated as previously described (4). Virus titers were determined by plaque assay as previously described (6).

Infection of mice and measurement of systemic disease

Mice were immunized with 3 × 10⁶ PFU of a recombinant vacv expressing β-galactosidase (vacvβ-gal) or vacvG and challenged with 2 - 3 × 10⁶ PFU of RSV as previously described (4). Weight loss and clinical illness scores were determined on a daily basis following RSV challenge as previously described (4).

Detection of eosinophils

Cells in the bronchioalveolar lavage (BAL) and lung were harvested as previously described (6). BAL and lung cells were counted and stained as previously described (6) with anti-Siglec-F PE, anti-CD11c APC, and anti-CD45 PE-Cy7. Cells staining positive for CD45 were gated and examined for Siglec-F and CD11c expression. A distinct population of cells staining Siglec-F positive and CD11c low were identified in vacvG-but not vacvβ-gal-immunized wild-type mice. These cells have previously been identified as eosinophils (7). The total number of eosinophils in the BAL and lung was calculated based on cell counts and the percents obtained from flow cytometry. Cells were collected on a FACSCanto (Becton Dickinson) and analyzed using Flow-Jo software (Tree Star).

Measurement of airway resistance

Enhanced pause (Penh) was measured using a whole body pleythsmograph (Buxco Electronics). Baseline Penh values for each mouse were recorded prior to and on a daily basis following RSV infection.

Statistics

Statistics analyses were performed using GraphPad InStat software.

Results and Discussion

STAT6 is required for the development of pulmonary eosinophilia

To evaluate the contribution of Th2 cells to RSV vaccine-enhanced disease, we examined the necessity of STAT6, a transcription factor activated after IL-4 and/or IL-13 cytokine stimulation that plays a critical role in Th2 cell differentiation (8). The number of eosinophils in the BAL and lung were determined 7 days after RSV challenge of mice previously immunized with either vacvG or a recombinant vacv expressing β-galactosidase (vacvβ-gal), as a control. As expected, vacvG-immunized wild-type (WT) mice undergoing a challenge RSV infection developed increased pulmonary eosinophilia in both the BAL and lung as compared to mice undergoing a primary RSV infection (vacvβ-gal-immunized). In contrast, vacvG-immunized STAT6-deficient mice had significantly decreased pulmonary eosinophila in both the BAL (Fig. 1A) and lung (Fig. 1B) after RSV challenge as compared to WT mice. As an additional control, mice deficient in STAT4 were also examined and were found to have no significant alteration in the development of pulmonary eosinophilia as compared to WT mice. These data support our previous work (4) and suggest that IL-13 signaling through STAT6 is necessary for the development of pulmonary eosinophilia in vacvG-immunized mice undergoing a challenge RSV infection.

Figure 1.

Pulmonary eosinophilia is dependent on STAT6. Cells in the BAL and lung of vacvβ-gal- and vacvG-immunized wild-type (WT), STAT4-, and STAT6-deficient (KO) mice were collected 7 days after RSV challenge. The total number of eosinophils in the BAL (A) and lung parenchyma (B) was calculated based on the percentage of CD45⁺ Siglec-F⁺ CD11C^low cells as determined by flow cytometry and the total cell counts. Data are shown from 2-4 mice per group and are representative of 5 independent experiments. Results are shown as mean ± SD. p < 0.01 (*) and p < 0.001 (**) compared to vacvG-immunized WT mice as determined by a Kruskal-Wallis test with Dunn post test.

STAT4 is necessary for increased weight loss, clinical illness, and Penh

To determine the contribution of Th2 cell-associated signaling through STAT6 to the development of RSV vaccine-enhanced systemic disease, weight loss and clinical illness scores were recorded on a daily basis after RSV challenge of STAT6-deficient mice previously immunized with vacvG. In addition, we examined vacvG-immunized STAT4-deficient mice to determine the relative contribution of Th1 cell-associated signaling to RSV vaccine-enhanced disease. Weight loss (Fig. 2A) and clinical illness scores (Fig. 2B) observed in STAT6-deficient mice were not significantly different from WT mice. However, as early as day 2 post-infection, STAT4-deficient mice exhibited significantly reduced weight loss and clinical illness as compared to WT mice. These data indicate that signaling through STAT4 is necessary for the development of RSV vaccine-enhanced systemic disease in vacvG-immunized mice undergoing a challenge RSV infection.

Figure 2.

Systemic disease and airway resistance are reduced in the absence of STAT4. Wild-type (WT), STAT4-, and STAT6-deficient (KO) mice previously immunized with vacvG were weighed (A) and assigned a clinical illness score (B) on a daily basis after RSV challenge. Results are presented as mean ± SD. A whole body pleythsmograph was used to measure airway resistance (C) on a daily basis after RSV challenge. The baseline enhanced pause (Penh) is shown as mean ± SEM. Data shown are from 4 mice per group and are representative of 5 independent experiments. p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***) between STAT4 KO and WT mice as determined by an ANOVA with Bonferroni post test.

Enhanced respiratory disease upon natural RSV infection was an important manifestation of the augmented illness observed in the children that received the FI-RSV vaccine. We used a whole body pleythsmograph (WPB) to determine if the increased Penh in vacvG-immunized mice undergoing a challenge RSV infection was dependent upon either STAT6 or STAT4. We verified the use of a WBP in this model by comparing enhanced pause (Penh) with resistance and elastance parameters obtained from intubated mice 3 days after RSV challenge (Supplementary Fig. S1). Fig. 2C demonstrates that increased Penh occurs as early as 1 day post-infection in both WT and STAT6-deficient mice previously immunized with vacvG. However, Penh is significantly decreased in STAT4-deficient mice as compared to WT mice. Together, these data demonstrate that signaling through STAT4 is necessary for the development of increased weight loss, clinical illness, and Penh upon RSV challenge of mice previously immunized with vacvG. It is noteworthy that the differences in weight loss, clinical illness, and Penh observed in the STAT4-deficient mice occur even though the magnitude of the cellular infiltrate in the lung in unaltered (Supplementary Fig. S2). Interestingly, because STAT6-deficient mice fail to develop pulmonary eosinophilia, our data also indicate that eosinophils do not contribute to several common measures of RSV vaccine-enhanced disease.

Eosinophils are not necessary for increased weight loss, clinical illness, or Penh

In order to more directly test whether or not eosinophils were necessary for weight loss, clinical illness, or increased Penh, we examined ΔdblGATA mice that cannot generate eosinophils from their bone marrow precursors due to the deletion of a palindromic binding site in the proximal promoter of the GATA-1 gene (5). Fig. 3A shows that RSV-induced weight loss is comparable between both vacvG-immunized ΔdblGATA and WT mice. In addition, ΔdblGATA mice exhibit increased clinical illness (Fig. 3B) and Penh (Fig. 3C) that is similar to that observed in WT mice. These data demonstrate that eosinophils are not required for increased weight loss, clinical illness, and Penh, the disease parameters most often associated with RSV vaccine-enhanced disease in this model.

Figure 3.

Eosinophils are not required for weight loss, clinical illness, or airway resistance. Wild-type (WT) and ΔdblGATA mice that were previously immunized with vacvG were weighed, assigned clinical illness scores, and analyzed for airway resistance on a daily basis after RSV challenge. Weight loss (A) and clinical illness scores (B) are presented as mean ± SD. Penh (C) was determined as described for Fig. 2 and is presented as mean ± SEM. Data shown are from 4 mice per group and are representative of 3 independent experiments. p < 0.05 (*) between ΔdblGATA and WT mice as determined by an unpaired t-test.

IFN-γ production contributes to clinical illness and increased Penh, but not weight loss

Activation of the transcription factor STAT4 augments IFN-γ production by Th1 cells (9). To determine the impact of IFN-γ production on RSV vaccine-enhanced disease, we examined weight loss, clinical illness, and Penh after RSV challenge of IFN-γ-deficient mice previously immunized with vacvG. Consistent with previous data demonstrating a requirement for TNF-α in mediating weight loss in this model (10), Fig. 4A demonstrates that in the absence of IFN-γ, weight loss is not significantly altered. However, both clinical illness (Fig. 4B) and Penh (Fig. 4C) were decreased in IFN-γ-deficient mice as compared to WT mice. However, since this reduction was not complete, these data indicate that IFN-γ contributes to, but is not the sole mediator of clinical illness and increased Penh in vacvG-immunized mice undergoing RSV challenge. Thus, it is likely that additional cytokines such as TNF-α may also contribute to the development of clinical illness and increased Penh in this model. Taken together, these data suggest that the activation of STAT4 leads to the production of IFN-γ and the subsequent development of clinical illness and increased Penh, as well as the production of TNF-α and the subsequent development of weight loss after RSV challenge of mice previously immunized with vacvG. Interestingly, depletion of IL-4 from FI-RSV-immunized mice has been shown to reduce weight loss upon RSV challenge (11). The cytokine requirements for the development of pulmonary eosinophilia have been found to differ between vacvG- and FI-RSV-immunized mice (4, 12). Thus, the cytokines responsible for inducing weight loss in each of these models may also differ.

Figure 4.

IFN-γ contributes to clinical illness and decreased pulmonary function, but not weight loss. Wild-type (WT) and IFN-γ-deficient (KO) mice that were previously immunized with vacvG were weighed, assigned clinical illness scores, and analyzed for airway resistance on a daily basis after RSV challenge. Weight loss (A) and clinical illness scores (B) are presented as mean ± SD. Penh (C) was determined as described for Fig. 2 and is presented as mean ± SEM. Data shown are from 4 mice per group and are representative of 3 independent experiments. p < 0.05 (*) and p < 0.01 (**) between IFN-γ KO and WT mice as determined by an unpaired t-test.

IL-12 and/or IL-23 cytokine signaling through their respective cellular receptors induces the activation of STAT4 (13, 14). IL-12 and IL-23 are both heterodimeric proteins that share that IL-12p40 subunit (15, 16). In order to determine if either IL-12 or IL-23 signaling through STAT4 was required for the development of RSV vaccine-enhanced disease, we examined weight loss, clinical illness, and Penh after RSV challenge of IL-12p40-deficient mice previously immunized with vacvG. We found none of these parameters significantly decreased in the absence of IL-12p40 as compared to WT mice (Supplementary Fig. S3). Therefore, neither IL-12 nor IL-23 signaling through STAT4 contributes to the development of RSV vaccine-enhanced disease suggesting that another STAT4-activating cytokine plays an important role in the induction of disease. One potential candidate is type 1 IFN, which has been demonstrated to activate STAT4 (17), and has been shown to be released in vivo after RSV infection (18).

The development of pulmonary eosinophilia has often been used as a hallmark of RSV vaccine-enhanced disease. However, the direct contribution of eosinophils to the development of RSV vaccine-enhanced disease has not been previously examined. Our data demonstrate that eosinophils do not contribute to multiple disease parameters including increased weight loss, clinical illness, and Penh upon RSV challenge of mice previously immunized with vacvG. The presence of eosinophils in the RSV infected lung appears to be primarily a consequence of the induction of a relatively low number of RSV-specific memory Th2 cells rather than a direct cause of significant pathology. Instead, our data suggests that signaling through STAT4 leads to increased weight loss, clinical illness, and Penh upon RSV challenge of mice previously immunized with vacvG.

The data presented here have important implications for future RSV vaccine design. Due to the previous notion that Th2 cells and eosinophils directly contributed to the development of RSV vaccine-enhanced disease, much effort in RSV vaccine design has gone into developing vaccination strategies to favor the induction of highly skewed RSV-specific Th1 responses. Our data demonstrate that eosinophils do not directly contribute to RSV vaccine-enhanced disease and indicate that RSV vaccine-enhanced disease is the result of an exuberant RSV-specific memory Th1 cell response. Importantly, it is not the magnitude of the overall inflammation, but the character of the immune response that dictates immunopathology. Thus, either skewing the RSV-specific memory CD4 T cell response to predominately a Th1 response or blocking the Th2 memory response all together would not prevent the development of RSV-induced disease but could, in fact, induce exacerbated disease. Our results suggest that future RSV immunization strategies should strive to achieve a balanced immune response.