Allergic pulmonary inflammation in mice is dependent on eosinophil-induced recruitment of effector T cells

We used a novel eosinophil-deficient transgenic line (PHIL), which is congenitally devoid of eosinophils without disturbing the production of other hematopoietic lineages (5), to unequivocally show the relationship between eosinophils and the development of pulmonary Th2 immune responses in a mouse model of acute respiratory inflammation. Specifically, PHIL mice and C57BL/6J wild-type controls were sensitized (days 0 and 14) with OVA/Alum and challenged with an OVA aerosol on days 24, 25, and 26 (control animals received saline alone), and endpoints of pathology were analyzed on day 28 (Fig. S1 A). In contrast to wild-type controls, PHIL mice were unable to induce a pulmonary eosinophilia (Fig. 1 A) or elevate bronchoalveolar lavage (BAL) Th2 cytokines (IL-4, IL-5, and IL-13) in response to allergen (Fig. 1 B). Previous hypotheses characterizing the immune responses in the lung after allergen provocation have implicated T cells in the production of Th2 cytokines (1). Furthermore, mouse models of chronic allergic asthma have correlated reduced airway eosinophilia in CCR3 knockout mice and eosinophil-deficient mice with reduced lymphocyte infiltration and hindered Th2 pathologies (12). This led us to determine the numbers of lymphocytes accumulating in the airways of allergen-provoked PHIL mice relative to wild-type animals as a possible explanation of the reduced Th2 responses in the absence of eosinophils. These assessments showed that lymphocyte numbers in the airways of OVA-treated PHIL mice were significantly reduced as compared with similarly treated wild-type controls (Fig. 1 C). Our studies also demonstrated that the decreased lymphocyte infiltration into the airways of PHIL mice was not transient and that the loss of eosinophils had consequences for the extended kinetics of lymphocyte accumulation. Specifically, PHIL and wild-type controls were allergen challenged and analyzed at different time points after allergen challenge (Fig. S1 B). BAL cell accumulation data demonstrated that the eosinophil infiltration of the airways in wild-type mice occurred maximally 4 d after allergen challenge and were resolved (i.e., returned to baseline levels) within 2 wk (Fig. 1 D). In contrast, although lymphocyte infiltration of the lungs in wild-type mice followed a similar kinetic pattern early (i.e., achieving a maximum at ∼4 d after challenge), these cells remained elevated during the subsequent 15-d period after allergen challenge (Fig. 1 E). However, these data also showed that in the absence of eosinophils, the kinetics of lymphocyte accumulation in the airways are disrupted with neither the maximal increase at 4 d after allergen challenge nor the sustained elevation of lymphocytes occurring in OVA-treated PHIL mice (Fig. 1 E). Thus, allergen-mediated eosinophil recruitment to the lungs of mice appears to be necessary for the pulmonary recruitment and accumulation of lymphocytes, and the corresponding production of BAL Th2 cytokines.

The lack of significant accumulation of total lymphocytes and the reduced Th2 cytokine expression suggested the potential that both CD4⁺ and CD8⁺ T cell accumulation in the lungs may be altered in OVA-treated PHIL mice (15). Leukocytes from the airways (i.e., BAL), whole-lung digests, caudal mediastinal lymph nodes, and splenocytes from OVA-sensitized/aerosol-challenged PHIL and wild-type animals were analyzed by flow cytometry for populations of CD4⁺ and CD8⁺ T cells (Fig. 2). In contrast to OVA-treated wild-type controls, which displayed significant increases in both T cell subtypes relative to saline control groups, OVA-treated PHIL mice failed to display allergen-mediated increases in either CD4⁺ or CD8⁺ T cells in the airways, lungs, and regional draining lymph nodes. This lack of T cell accumulation was restricted to the lungs and lymph nodes, as spleens from these mice did not statistically differ in their total numbers of CD4⁺ and CD8⁺ T cells, suggesting an inability to accumulate these cells specifically to pulmonary compartments after OVA provocation.

Mouse models of allergen-mediated Th2 inflammation have been attributed to the activities (e.g., Th2 cytokine expression) of particular subsets of CD4⁺ T cells in the lung (e.g., CD4⁺/CD25⁺ cells [i.e., proliferating T cells and/or regulatory T cells; reference 16], NKT cells [CD4⁺/NK1.1⁺; reference 17]), and effector T cells (CD4⁺/CD62L^−/lo/CD44^hi [18]), as well as CD8⁺ T cells (19). To define the extent of the effects on pulmonary T cell recruitment in the absence of eosinophils, the numbers of CD4⁺ T cells within each subset were determined in OVA-treated PHIL mice relative to their accumulation in wild-type control animals. The lungs of OVA-treated PHIL mice displayed a significant reduction in accumulating CD4⁺/CD25⁺ T cells and CD4⁺/NK1.1⁺ NKT cells (Fig. 3 A) relative to OVA-treated wild-type controls. Furthermore, examination of the T cell subtypes in OVA-treated PHIL versus wild-type mice showed that significant accumulation of effector T cells (CD4⁺/CD62L^−/lo/CD44^hi) in the lungs and regional draining lymph nodes occurred only in wild-type animals and was absent in OVA-treated PHIL mice (Fig. 3 B). Interestingly, the regional draining lymph nodes of OVA-treated PHIL mice also displayed a significant reduction in memory (CD4⁺/CD62L^med/hi/CD44^hi) but not naive (CD4⁺/CD62L^hi/CD44^−/lo) T cells (Fig. 3 B). Furthermore, assessments of CCR7 expression (i.e., a memory/effector T cell marker that appears to participate in the recruitment of T cells to lymph nodes in allergen challenge models [20]) also demonstrated that CD4⁺/CCR7⁺ T cells were reduced in OVA-treated PHIL draining lymph nodes as compared with wild-type controls (0.65 ± 0.12 vs. 2.18 ± 0.35 × 10⁵, respectively; P < 0.05). These phenomena appear to be specific to the pulmonary compartments, as splenocyte numbers of effector, memory, and naive T cells are unchanged between OVA-treated PHIL and wild-type control animals. Thus, the loss of these T cell populations in the lungs and regional draining lymph nodes of OVA-treated PHIL mice suggested that the absence of eosinophils led to the loss of T cell proliferation/activation, a loss of recruitment pathways leading to T cell accumulation in the lung, or both.

Several previous studies have demonstrated that eosinophils participate in the proliferation/activation of naive and memory T cells into effector T cells (e.g., through antigen presentation and co-stimulation [14, 21–23]), which may be evident by the reduced numbers of memory/effector T cells in the regional draining lymph nodes of OVA-treated PHIL mice. Conversely, the impaired recruitment of all T cell subsets (CD8⁺, CD4⁺/CD25⁺, CD4⁺/NK1.1⁺, and CD4⁺/CD62L^−/lo/CD44^hi) into the lung may implicate a larger and more profound failure in T cell recruitment, extending beyond potential eosinophil activities mediating allergen-specific T cell proliferation/activation. We assessed the potential of each of these eosinophil-dependent mechanisms as contributors to pulmonary T cell accumulation in two independent experiments.

First, in eosinophil-mediated T cell proliferation/activation, our inability to recruit OVA-specific Th2 cells after the adoptive transfer of eosinophils into the airways of OVA-sensitized and -challenged PHIL mice (unpublished data) suggested that eosinophils may be required for Th2 cell generation. This potential pathway was further defined in ex vivo studies demonstrating that eosinophils were required for the elaboration of Th2 immune responses after allergen challenge. Specifically, splenocytes from OVA-sensitized or OVA-sensitized and -challenged PHIL mice (splenocytes from wild-type animals were used as controls) were isolated, and their abilities to elaborate IL-13 were assessed after co-culture with media alone, OVA, or a nonspecific stimulus (i.e., cross-linking of surface receptors with anti-CD3 and -CD28 antibodies; Fig. 3 C). These studies demonstrated that the absence of eosinophils (i.e., PHIL mice) leads to a defect in OVA-specific Th2 cytokine production. That is, this observed defect was restricted to OVA exposure ex vivo, as T cell responses to a nonspecific stimulus were equivalent between splenocytes isolated from PHIL versus wild-type mice. Significantly, this OVA-specific defect in IL-13 production was limited to animals that were both OVA-sensitized and aerosol-challenged, suggesting an impairment in the activation/proliferation of OVA-specific memory T cells into polarized Th2 effector T cells in OVA-treated PHIL mice. This immunoregulatory role for eosinophils represents a likely mechanism contributing to the attenuated immune responses in OVA-treated PHIL mice and is part of our ongoing studies further characterizing the role of eosinophils in activation/co-stimulation of T cells (unpublished data).

Second, in eosinophil-dependent recruitment/accumulation of T cells to the lung, the impaired accumulation of multiple T cell subsets in OVA-treated PHIL mice (i.e., CD8⁺, CD4⁺/CD25⁺, CD4⁺/NK1.1⁺, and CD4⁺/CD62L^−/lo/CD44^hi T cells) suggested that eosinophils may be required for generalized pathways that promote localized T cell recruitment. The ability of all of these T cell subtypes to express CCR4 (24–26) also provides a potential mechanism by which eosinophils would be able to mediate the recruitment of T cells to sites of Th2 inflammation. Several lines of evidence further suggested that the CCR4 chemokine ligands TARC and MDC were likely candidates by which lung-infiltrating eosinophils may elicit the recruitment of T cells: (a) TARC and MDC have been described as part of the mechanisms mediating T cell recruitment in several mouse models of Th2 inflammation (26); (b) induced airway epithelial and macrophage expression of TARC and MDC correlates with the pulmonary eosinophilia associated with allergic respiratory inflammation (27), and the levels of each of these chemokines are elevated in the sputum of asthmatic patients as compared with otherwise normal individuals (28); and (c) microarray analyses of gene expression comparing mRNA levels from BAL-derived eosinophils of OVA-treated mice versus blood eosinophils of OVA-naive animals showed that TARC and MDC were each significantly elevated (greater than fourfold) in BAL-derived eosinophils (unpublished data) (29).

Assessments of MDC and TARC levels in the BAL of OVA-treated PHIL and wild-type mice confirmed that the expression of these chemokines was linked to the presence of pulmonary eosinophils. That is, OVA-sensitized/aerosol-challenged wild-type animals displayed a significant elevation (relative to saline controls) in BAL levels of MDC and TARC that did not occur in OVA-treated PHIL mice (Fig. 3 D). In contrast, the BAL levels of other chemokines such as eotaxin-1 and -2 were not statistically different between PHIL and wild-type mice (not depicted). Kinetic assessments of MDC and TARC expression in OVA-treated wild-type animals showed that this expression correlated with the induced pulmonary eosinophil infiltrate, with BAL chemokine levels closely paralleling BAL eosinophilia after OVA challenge (Figs. 3 E). Significantly, the levels of each of these Th2 cell chemokines in OVA-treated PHIL mice were indistinguishable from saline controls throughout the time period examined, suggesting that eosinophil-induced expression of MDC and TARC may be a contributing (possibly responsible) mechanism for allergen-mediated T cell recruitment to the lung.

Adoptive cell transfers of OVA-specific Th2-polarized T cells into OVA-challenged PHIL recipient mice were performed to demonstrate the role of eosinophils in T cell recruitment and the development of Th2-driven immune responses in the lung. Specifically, adoptive transfer of OVA-specific Th2-polarized effector T cells allowed us to bypass requirements of eosinophil-dependent mechanisms of T cell proliferation/activation, testing the hypothesis that eosinophils contribute to the development of allergen-mediated Th2 immune responses through the recruitment of OVA-specific effector T cells. These studies capitalized both on our previous experience with eosinophil adoptive transfer (30) as well as the availability of OVA-specific TCR transgenic mice that allows the isolation/transfer of in vitro Th2-polarized OVA-specific effector T cells (OT-II). In this allergen challenge model system, naive eosinophil-sufficient wild-type and eosinophil-deficient PHIL recipient mice were either injected (i.v.) once with PBS or 1.5 × 10⁷ OT-II cells on day 0 of this protocol, followed by intranasal (i.n.) administration of 20 μl of a 1 μg/μl OVA solution prepared in saline on protocol days 0, 1, and 2 (Fig. 4 A). OVA-induced responses in these recipient mice were assessed on day 4 as the elaboration of Th2 cytokine expression and/or the development of allergen-induced pulmonary histopathology. Because of the eosinophil-deficient character of PHIL recipient mice and, therefore, the potential of artifactual events associated with the adoptive transfer of Th2-polarized OVA-specific effector T cells into these recipients, several control studies were performed to ensure that the proliferation and cell-surface activation profile of the adoptively transferred OT-II cells were not altered relative to wild-type control mice. For example, OT-II cells were labeled with the fluorescent marker CFSE, and the proliferation and cell-surface activation markers of these cells were assessed in OVA-challenged PHIL versus wild-type mice after adoptive transfer. As shown in Fig. S2 , OT-II cell proliferation (as judged by the dilution of CFSE staining intensity) was similar between PHIL and wild-type mice, indicating that eosinophils were not required for in vivo proliferation of these T cells. In addition, these CFSE-labeled cells maintained their memory (CD4⁺/CD62L^med/hi/CD44^hi)/effector (CD4⁺/CD62L^−/lo/CD44^hi) T cell surface phenotypes in OVA-challenged PHIL animals after adoptive transfer (Fig. S3).

A total of three independent experimental groups of OVA-challenged animals was examined using this adoptive transfer model to define the relative roles of eosinophils in the development of Th2 pulmonary immune/inflammatory responses (Fig. 4 A). Group I consisted of OVA-naive wild-type and PHIL recipient mice receiving OT-II cells i.v. (controls received PBS vehicle alone) on protocol day 0, and saline intratracheally (i.t.) on protocol days 0, 1, 2, and 3. Group II consisted of OVA-naive PHIL recipient mice after adoptive transfer (i.t.) of 10⁷ eosinophils on protocol days 0, 1, 2, and 3. The success of eosinophil adoptive transfer was confirmed by the recovery of BAL eosinophils in hematoxylin and eosin–stained cytospin preparations (Fig. S4 A) and the identification of tissue-infiltrating eosinophils using major basic protein (MBP)–specific immunohistochemistry (Fig. S4 B) on protocol day 4. Finally, group III consisted of OVA-naive PHIL recipient mice receiving both OT-II cells (i.v.) on protocol day 0 and adoptively transferred (i.t.) eosinophils on protocol days 0, 1, 2, and 3.

These studies showed that the adoptive transfer of OT-II cells (group I) or eosinophils (group II) alone into OVA-challenged PHIL recipient mice were each insufficient for the development of allergen-induced Th2 pulmonary responses (Fig. 4 B). Moreover, OVA-naive wild-type and PHIL mice receiving OT-II cells by i.v. and saline by i.n. challenge did not display an increase in either Th2 cytokine levels or induced histopathologies (not depicted). In contrast, adoptive transfer of both OT-II cells and eosinophils into PHIL mice (group III) led to significant increases in BAL levels of the Th2 cytokines (e.g., IL-5 and IL-13; Fig. 4 B); IL-4 was not significantly elevated above baseline. The development of Th2-driven histopathologies in each of these groups of mice paralleled the pattern of induced Th2 cytokine expression. That is, administration of either OT-II cells (group I) or eosinophils (group II) alone into OVA-challenged PHIL mice failed to result in the development of histopathologies such as the peribronchial aggregation of eosinophils (Fig. S4 B) and goblet cell metaplasia/airway epithelial cell mucin accumulation (Fig. 4 C). However, the adoptive transfer of both OT-II cells and eosinophils into OVA-challenged PHIL mice (group III) resulted in goblet cell metaplasia/airway epithelial cell mucin accumulation that was comparable to that observed in OVA-challenged wild-type mice receiving OT-II cells (Fig. 4 C). These observations suggest that the eosinophil activities required for the development of Th2 responses extend beyond the potential effects on the proliferation/activation of OVA-specific effector T cells.

The allergen-induced pulmonary pathologies resulting from the combined adoptive transfer of OT-II cells and eosinophils (group III) suggested that eosinophil-dependent recruitment of T cells may be a responsible mechanism for the immune/inflammatory responses that occur in the lungs of allergen-challenged mice. Indeed, unlike the transfer of individual leukocyte populations (group I and II), the concurrent adoptive transfer of both OT-II cells and eosinophils (group III) resulted in a significant increase in the recruitment of CD4⁺ T cells in the airways and lung tissue of OVA-challenged PHIL mice (Fig. 5 A); CD8⁺, CD4⁺/CD25⁺, and CD4⁺/NK1.1⁺ T cell profiles were similar to CD4⁺ T cell profiles (not depicted). Thus, adoptive transfer of eosinophils in the presence of OVA-specific Th2 effector T cells restored the recruitment of T cell subsets to the lung concurrent with the observed increases in Th2 immune-mediated pathologies.

To demonstrate that the recruitment of T cells to the lungs of OVA-challenged mice after adoptive cell transfer correlated with MDC and TARC expression, we measured the BAL levels of these chemokines in each of the adoptive transfer groups of mice. Significantly, the expression of MDC and TARC in the BAL of mice after adoptive cell transfer paralleled the pattern of CD4⁺ T cell accumulation observed in these mice (Fig. 5 B). That is, PHIL mice receiving only OT-II cells (group I) or eosinophils (group II) were unable to elevate MDC and TARC to the levels observed in OVA-challenged wild-type mice receiving OT-II cells, whereas the administration of both OT-II cells and eosinophils to OVA-challenged PHIL mice was sufficient to return these chemokines to wild-type levels.

We performed antibody blocking studies to antagonize the activities of MDC and TARC after adoptive cell transfer to demonstrate that the recruitment of T cells in group III mice was dependent on the observed eosinophil-induced expression of these Th2 chemokines. Specifically, in addition to the assessments of CD4⁺ T cell recruitment to the lung after adoptive transfer of either OT-II cells (group I) or eosinophils (group II) alone, OVA-challenged PHIL mice adoptively transferred with both OT-II cells and eosinophils (group III) were treated with either isotype control antibody or neutralizing antibodies to TARC and MDC on days 0, 1, 2, and 3 of the protocol outlined in Fig. 4 A and Materials and methods. The antibody-mediated blockade of TARC and MDC in group III mice returned CD4⁺ T cell recruitment in these PHIL animals to the baseline levels observed in PHIL mice from groups I and II (Fig. 5 C). These observations extended to multiple T cell subsets (e.g., CD4⁺/CD62L^−/lo/CD44^hi effector T cells and CD8⁺ T cells) that were also dramatically reduced in the lungs of anti-TARC/-MDC–treated group III mice (not depicted). These studies showed that T cells are recruited to the lung by the eosinophil-dependent production of TARC and MDC and that this induced expression after allergen challenge is mechanistically linked to the T cell recruitment occurring in these mice.

The observation that eosinophils accumulating in the lungs after allergen provocation modulate local tissue immune responses through the specific recruitment of T cells suggests a novel mechanism of eosinophil effector function whose significance was previously underappreciated. This mechanism represents an expansion/extension of earlier studies investigating the potential importance of eosinophils and also provides an alternative interpretation of data initially noted by Lloyd et al. (26). Specifically, we suggest that the expression of eotaxins and perhaps other eosinophil agonist factors (e.g., leukotriene B₄) (31) immediately after allergen challenge elicits the recruitment of eosinophils and the expression of cytokines and growth factors that also promote eosinophil survival and activation. In turn, if and when these accumulating eosinophils achieve a critical level, their activities promote the local expression of Th2 chemokines (e.g., TARC and MDC) that leads to and/or amplifies the recruitment of T cells to the lungs and allergen-mediated Th2 inflammation. Thus, either by a direct (i.e., cell autonomous) or an indirect (i.e., induced from a third party cell type) mechanism, infiltrating lung eosinophils may mediate the induction of TARC and MDC that elicits the T cell recruitment necessary for the inflammatory responses associated with mouse models of asthma. Interestingly, studies assessing the initial leukocyte recruitment to the lungs in both human subjects and mouse models of allergic respiratory inflammation demonstrated similar kinetic patterns of eosinophil and lymphocyte accumulation into the airways (32, 33), implying that eosinophil-dependent accumulation of lymphocytes leading to pulmonary Th2 inflammatory responses may represent a larger, more significant mechanism with implications for asthma patients.

Our larger hypothesis linking eosinophils, T cells, and allergic respiratory inflammation is that eosinophils mediate the recruitment of effector T cells by two apparently independent mechanisms: (a) activation and co-stimulation of T cells promoting proliferation and/or increased survival, and (b) pulmonary production of chemokines that promote the recruitment and/or accumulation of T cells in the lung. It is also noteworthy that these eosinophil-dependent mechanisms are not mutually exclusive, and each may contribute to the loss of Th2 immune responses and pulmonary pathologies in OVA-treated PHIL mice. We suggest that these dual functions, as well as their requirement for a threshold level of eosinophils, are responsible for many of the difficulties associated with the interpretation of results from both earlier mouse-model studies of allergic inflammation and clinical studies of asthma patients that target eosinophils. For example, the partial eosinophil deficiency of IL-5^−/− mice after OVA sensitization/challenge (∼90% reduction relative to wild-type mice) is likely only to attenuate and not abolish eosinophil-mediated effects on T cells. This, in part, may explain the differences in allergen-induced pulmonary inflammation that occurred in IL-5^−/− mice among different studies (e.g., reference 34 vs. reference 35). Indeed, a requirement for a threshold level of recruited eosinophils was suggested in studies examining the allergic respiratory inflammation in OVA-treated IL-5^−/− mice relative to OVA-treated IL-5/eotaxin-1 double-knockout (IL-5^−/−/eotaxin-1^−/−) animals. That is, although the partial loss of eosinophils failed to prevent OVA-induced inflammatory responses in IL-5^−/−, these responses were abolished in IL-5^−/−/eotaxin-1^−/− mice that had a complete ablation of pulmonary eosinophils (36). The demonstrated inability to completely eliminate lung tissue eosinophils among human subjects receiving anti–IL-5 therapeutics may also represent a confounding issue that has complicated conclusions from clinical studies of asthma patients (6, 37, 38). These observations suggest that together with potential activities on T cell proliferation/activation and destructive effector activities mediated by granule protein release, the additional capability of eosinophils to mediate effector T cell accumulation in the lung may provide yet another mechanism by which eosinophils initiate and maintain the immune/inflammatory responses associated with allergic asthma. Interestingly, two unrelated studies (12, 39) using another eosinophil-deficient strain of mice (i.e., ΔdblGATA-1 [40]) also provide relevant insights supporting the immunoregulative roles of eosinophils. Fulkerson et al. (12) reported that microarray assessments of lung gene expression in Aspergillus allergen-challenged ΔdblGATA-1 mice (relative to allergen-challenged wild-type animals) showed that the absence of eosinophils was associated with a significant reduction in pulmonary Th2 gene expression (e.g., IL-4 and IL-13). In addition, Voehringer et al. (39), using ΔdblGATA eosinophil-deficient mice and a Nocardia brasiliensis mouse model of lung inflammation, concluded that eosinophils were not required for pulmonary T cell recruitment during the primary infection of mice. However, eosinophils in this study may be linked with a failure to activate/recruit CD4⁺ T cells to the lung and, in turn, inflammatory responses associated with secondary challenges with this parasite. Thus, data from both of these studies highlight a complexity of pulmonary immune responses but again imply a potentially dual role for eosinophils associated with the elaboration of secondary immune responses in the lung, leading to the proliferation/activation of antigen-specific T cells and/or subsequent recruitment of these T cells to the lung.

Finally, there is no reason to assume that either of the eosinophil-mediated effects on T cells outlined in this paper are necessarily mutually exclusive, and each is likely to contribute to the induced pulmonary immune responses observed in eosinophil-sufficient wild-type mice. These eosinophil-dependent immunoregulatory mechanisms are also not likely to be limited to the lung and/or allergen-mediated inflammation. That is, the accumulation of tissue eosinophils in a variety of settings may represent provocative examples in which eosinophil-mediated expression of Th2 chemokines leads to the recruitment of effector T cell recruitment and the modulation of local tissue immune cascades. In particular, we suggest that the T cell–induced immune responses associated with acute rejection of transplanted organs (41), the onset/growth of tumors (42), the accumulation/differentiation of thymic T cell populations (43), and the immune modulation of the gastrointestinal tract mucosa (44) may ultimately each be causatively linked to tissue-specific recruitment of activated effector T cells by eosinophils accumulating at these sites.

All studies were performed with either C57BL/6J wild-type or eosinophil-deficient PHIL mice (5) crossed to C57BL/6J for >10 generations. C57BL/6J mice and TCR OVA-transgenic animals on a C57BL/6J background (OT-II) were purchased from the Jackson Laboratory. The mice were maintained in ventilated microisolator cages housed in the specific pathogen-free animal facility at the Mayo Clinic Arizona. The sentinel cages within the animal colony surveyed negative for the presence of known mouse pathogens. Protocols and studies involving animals were performed in accordance with National Institutes of Health and Mayo Foundation institutional guidelines.

Mice were sensitized and challenged with OVA as previously described (5) and outlined in Fig. S1 A. The kinetics of leukocyte recruitment to the lungs of sensitized and challenged mice were performed as previously described (33) and outlined in Fig. S1 B. Studies involving adoptive transfer of T cells derived from OVA-specific TCR transgenic mice (OT-II cells) were performed using a modification of a protocol previously described (45) and outlined in Fig. 4 A. Specifically, on day 0 of this protocol, either PBS or 1.5 × 10⁷ in vitro Th2-polarized effector T cells (OT-II) were injected i.v. into OVA-naive wild-type or PHIL mice. Within 3 h of i.v. OT-II cell transfer, the mice were gently anesthetized and received an i.n. administration of 20 μg OVA (1 μg/μl in saline). Mice were additionally challenged by i.n. administration of OVA on days 1 and 2 of this protocol (i.e., mice were OVA challenged via the nasal passages on three consecutive days). In some groups of mice, 10⁷ eosinophils in 20 μl of saline were adoptively transferred (control animals received saline vehicle alone) via the trachea (i.t.) on days 0, 1, 2, and 3 of this protocol, and within 1 h after OVA administration on the allergen-challenge days. OVA-induced responses in these recipient mice were assessed on day 4.

BAL assessments were completed as described previously (5, 46).

Histopathologic changes of the airways were assessed as described previously (5). 4-μm formalin-fixed, paraffin-embedded sections of mouse lungs were stained for epithelial cell mucin accumulation using periodic acid Schiff or, for eosinophils, rabbit polyclonal anti–mouse eosinophil MBP antiserum (46).

Total pulmonary leukocytes were recovered from lungs perfused with PBS-EDTA before removal, placed in complete media with 175 U/ml collagenase IV, and digested for 45 min at 37°C, as described previously (47). Single-cell suspensions of lung, spleens, and caudal mediastinal lymph nodes (regional draining lymph nodes) were obtained by homogenization with frosted glass slides followed by passing through a 40-μm nylon filter to remove larger aggregates of cells per tissue. Red blood cells were lysed by ammonium chloride (PharmLyse; BD Biosciences). Viability of the cells was >95%, as determined by Trypan blue exclusion. Cell counts were completed after the last wash before staining for flow cytometry.

Mouse IL-4, IL-5, IL-13, MDC, and TARC levels were assessed using immunoassay kits (R&D Systems) according to the manufacturer's instructions. The limits of detection for each cytokine assay were 5–10 pg/ml.

Single-cell suspensions were obtained from sensitized and sensitized/challenged PHIL and wild-type mice using the protocol shown in Fig. S1 A. 10⁶ cells/well of splenocytes were cultured in a 96-well plate with either media alone, with 200 μg/ml OVA (Sigma-Aldrich), or with plates precoated with anti-CD3 and -CD28 (BD Biosciences) for 96 h. Supernatants were analyzed by ELISA.

Single-cell suspensions were stained for 25 min on ice with cell type–specific antibodies after blockade of Fc receptors using 1 μg/μl of Fc blocker (CD16/32; BD Biosciences). Antibodies used for staining specific cell types were obtained from BD Biosciences and include the T cell subtype–specific antibodies CD3, CD4, CD8, and CD62L, and the NK/NKT cell–specific marker NK1.1. Additional antibodies representing other T cell markers (i.e., CD44 and CD25) were purchased from eBioscience. Propidium iodide staining was used to identify and “gate out” dead cells. Flow cytometry was performed on a cytofluorimeter (FACScan; Becton Dickinson). Data acquisition and analysis were performed using CellQuest Pro (Becton Dickinson) and Sumit (version 4.3; Dako) software.

Naive OT-II cells were isolated from splenocytes and lymph nodes of 5–12-wk-old mice by two rounds of positive selection using magnetic beads coupled to the T cell–specific marker CD90⁺ (Thy1.2; MACS [Miltenyi Biotec]). 0.5 × 10⁶ purified CD90⁺ T cells/ml were differentiated to the Th2 effector phenotype by incubating the cells in RPMI 1640 media (supplemented with 15% FBS [HyClone Technologies], 2 × 10⁻⁵ M 2-mercaptoethanol, 1 mM sodium pyruvate, 1× nonessential amino acids, 10 μg/ml penicillin, 10 μg/ml streptomycin, and 2 mM l-glutamine; all from Invitrogen) in the presence of 50 μg/ml OVA 323-339 peptide, 8 ng/ml mouse IL-4 (R&D Systems), 5 μg/ml anti–IL-12 mAb (C17.8; BD Biosciences), and 5 μg/ml anti–IFN-γ (XMG1.2; BD Biosciences). Irradiated (30 Gy) splenic APCs (depleted of CD 90⁺ cells) were also added to these cultures (2 × 10⁶ cells/ml). After 2 d of culture, cells were given 10 ng/ml IL-2 (R&D Systems), and on days 4 and 6 cells were given fresh media with 10 ng/ml of IL-2. After culture for a total of 7 d, the recovered Th2-polarized effector T cells were used for i.v. injection (1.5 × 10⁷ cells/mouse) into recipient mice. In some experiments, cells were labeled with CFSE, as described by the manufacturer (Invitrogen), before injection. FACS analysis reveals >98% of cells using this method are CD3⁺, with a composition of ∼85% CD4⁺ and ∼15% CD8⁺. Trypan blue exclusion demonstrated >98% cell viability. Th2 polarization of this cell population was confirmed by ELISA assessments for the presence of the Th2 cytokines (i.e., IL-4, IL-5, and IL-13) of 7-d cell-culture supernatants.

Eosinophils were isolated from IL-5–expressing transgenic mice (NJ.1638) (30), as previously described. In brief, peripheral blood pooled from four mice was layered onto single-step Percoll gradients. Eosinophils from the buffy coat at the blood–Percoll interface were subsequently isolated by removing the contaminating lymphocytes using MACS with magnetic beads conjugated with CD45R/B220 (B cells) and CD90/Thy1.2 (T cells), according to the manufacturer's recommendations. Visual examination of DiffQuick-stained (Dade Behring Inc.) cytospin preparations and by flow cytometry analysis with staining for anti-CCR3 (R&D Systems) revealed that the purity of these eosinophil preparations was >99%.

The adoptive transfer of eosinophils and/or vehicle control (saline) into the lungs of mice was accomplished in subject animals lightly anesthetized with 1% isofluorane before i.t. instillation, as described previously (30). The transfer of eosinophils (or saline vehicle) was completed 1 h after i.n. administration of OVA on days 0, 1, and 2, and at the same time of day without OVA on protocol day 3 (Fig. 4 A).

Mice were treated with antibodies to MDC and TARC with minor modifications of a protocol described previously (48). Specifically, 1 h before the OVA challenges, the PHIL mice received an i.p. injection of either control polyclonal goat IgG antibody or 20 μg each of polyclonal IgG goat anti-TARC and anti-MDC (R&D Systems) on days 0, 1, 2, and 3 of the protocol outlined in Fig. 4 A.

All data were derived from at least three independent experiments, each with cohort sizes of two to five mice. Data presented are the means ± SE. Statistical analyses were performed using the Student's t test with differences between means considered significant when P < 0.05.

Fig. S1 presents schematic timelines of the acute OVA sensitization/aerosol challenge protocols used to elicit allergic respiratory inflammation, including the recruitment/accumulation of eosinophils and lymphocytes to the lung. Fig. S2 demonstrates that OT-II cells displayed the same proliferative capacity after adoptive transfer into eosinophil-sufficient wild-type and PHIL mice. Fig. S3 shows that OT-II cells displayed the same T cell surface phenotype after adoptive transfer into eosinophil-sufficient wild-type and PHIL mice. Fig. S4 demonstrates that OVA provocation of PHIL mice after adoptive transfer of both OT-II cells and eosinophils resulted in the development of a BAL eosinophilia and proinflammatory aggregates in the peribronchial regions of the lung.

The authors wish to thank the invaluable contribution of numerous individuals not listed as authors, including Teresa Tinder, Katie O'Neill, Alfred Doyle, and Dr. Michael McGarry. Moreover, we also wish to acknowledge the tireless efforts of the Mayo Clinic Arizona medical graphic artist Marv Ruona, Nichole Boruff, and Joseph Esposito of Research Library Services. In addition, we wish to express our gratitude to the Lee Laboratories administrative staff (Linda Mardel and Margaret McGarry), without whom we could not function as an integrated group.

The work presented was supported by the Mayo Foundation and grants from the National Institutes of Health to J.J. Lee (HL065228 and K26-RR019709), N.A. Lee (HL058723), and E.A. Jacobsen (HL08514). Additional support was provided by grants from the American Heart Association to J.J. Lee (045580Z) and N.A. Lee (0555639Z).

The authors have no conflicting financial interests.