Complex Role of the IL-4 Receptor α in a Murine Model of Airway Inflammation: Expression of the IL-4 Receptor α on Nonlymphoid Cells of Bone Marrow Origin Contributes to Severity of Inflammation¹

Asthma is a highly complex clinical disorder characterized by episodes of airway narrowing, inflammation, and excessive mucus production (1). A number of laboratories have clearly established the importance of the TH2-derived cytokines IL-4, IL-5, and IL-13 in mediating the airway inflammatory response using a murine model for allergic lung inflammation (2, 3, 4, 5, 6, 7, 8, 9, 10). The cytokines IL-4 and IL-13 share a common receptor complex and activate similar signal transduction pathways. In lymphoid cells, the type I receptor complex consists of the high affinity binding chain (IL-4Rα) (3) and the common γ-chain. The type II IL-4R complex binds both IL-4 and IL-13 and consists of the IL-4Rα and the IL-13Rα1 chain (11). The vast majority of cell types in the body express the type I or the type II (or both) IL-4R complex, including cells of the lung, brain, muscle, kidney, placenta, epithelium, and endothelium. Furthermore, most cells of bone marrow (BM)³ origin, including monocytes and eosinophils, express both receptor complexes (12, 13, 14, 15, 16, 17, 18). Where analyzed, the receptor-positive cells responded to IL-4 and IL-13 (19, 20, 21). However, the physiologic significance of the wide distribution of the receptor is not clear.

Even though the receptors for IL-4 and IL-13 are ubiquitously expressed, we propose their contribution to the pathogenesis of asthma may be cell type and context specific. Several laboratories have investigated the role IL-4Rα expression plays in regulating inflammatory responses using IL-4Rα knockout (KO) mice. Mice completely lacking the IL-4Rα do not develop allergic airway hyperreactivity, airway mucus production, or inflammation (9). In these mice, the development of TH2 cells and the cytokines they produce with the potential to influence the asthma phenotype were diminished (22). However, the unresponsiveness to Ag challenge observed in IL-4RαKO mice is not simply due to a deficiency in TH2 cytokine production. Grunig et al. (5) reported that intranasal treatment of IL-4RαKO mice with either IL-4 or IL-13 was not sufficient to induce airway hyperresponsiveness (AHR), goblet cell differentiation, or airway eosinophilia while such treatment was sufficient in wild-type mice. In addition, Cohn et al. (9) showed that transferring OVA-specific TH2 cells to IL-4RαKO mice was not sufficient to induce goblet cell differentiation or airway eosinophilia in response to OVA inhalation. These data suggested that the IL-4Rα on non-T cells, possibly on airway epithelial cells, participates in the development of airway inflammation and mucus production. Although suggestive, these studies did not conclusively demonstrate the relative contribution of other cell types expressing the IL-4Rα, including other BM-derived cells.

To characterize the role the IL-4Rα plays in the development of the characteristic features of allergic inflammation, we used a BM transfer approach that allowed us to specifically analyze the relative contribution of IL-4Rα expression on BM cells vs non-BM cells in a murine model of airway inflammation. We found that in the absence of IL-4Rα in recipient mice, there was no goblet cell metaplasia or mucus hypersecretion, even in the presence of IL-4Rα⁺ BM cells, TH2 cells, and substantial eosinophilic inflammation. Furthermore, we found that while the TH2 response is necessary, expression of the IL-4Rα on an additional MHC class II⁺, BM-derived cell type contributes to the severity of both inflammatory cell infiltrate and mucus production. Taken together, these results indicate that signaling through the IL-4Rα by IL-4 and/or IL-13 regulates the development of allergic inflammation at multiple levels.

BALB/c wild-type and RAG2-deficient mice (RAG2KO) were purchased from Taconic Farms (Germantown, NY). IL-4RαKO (on a BALB/c background) (22) and IL-4Rα × RAG2KO (23) were bred in the animal care facility at the National Institutes of Health (Rockville, MD) or at the American Red Cross. Bone marrow (BM) was isolated from both femur and tibia. Recipient mice were irradiated (500 rad) and BM was injected into recipients at a ratio of one donor to two recipients. The phenotypes of mice generated by BM transplants are shown in Table I. Six weeks was allowed for reconstitution and, where appropriate, reconstitution was analyzed by staining of peripheral blood cells for B and T cell markers. Five mice were routinely used for each condition studied in an individual experiment. Each transfer group and treatment condition was repeated at least three times. All experimental methods described in this manuscript were performed as approved by the Institutional Animal Care and Use Committee at the American Red Cross.

Mesenteric lymph nodes were isolated from DO11.10 × RAG2KO mice and differentiated to TH2 cells as previously described (24). Briefly, lymph node (LN) cells were stimulated with OVA peptide (1 μM) in the presence of IL-2 (20 ng/ml) (R&D Systems, Minneapolis, MN), IL-4 (50 ng/ml) (R&D Systems), anti-IFN-γ and anti-IL-12 (both from BD Biosciences/PharMingen, San Diego, CA) (10 μg/ml each) and a 20-fold excess of irradiated, BALB/c spleen cells. After 3 days, cells were expanded in the presence of IL-2 alone for 3–4 days. Subsequently, the cells were washed and treated with the same conditions for a second round of priming. The cytokine-producing phenotype of these cells was tested using PE-anti-IL-4, FITC-anti-IFN-γ, and an intracellular cytoplasmic staining kit (from BD PharMingen) as previously described (24). The TH2 cells were routinely ∼30–40% IL-4 producers and 0% IFN-γ producers. After the second round of priming, the cells were rested for 4 days, washed, and 2 × 10⁷ cells were injected i.v. via tail vein 1 day before immunization.

Mice were sensitized to chicken egg OVA (Sigma-Aldrich, St. Louis, MO) using a modified method as described by Hamelmann and Gelfand (25). OVA was adsorbed to aluminum hydroxide (Alum) as described (26). Mice were immunized with Alum alone or 100 μg of OVA/Alum via i.p. injection on day 1 and boosted with the same reagents on day 14. Mice were exposed to 1% OVA in PBS by nebulization for 20 min each day on days 19, 21, and 26.

Bronchial lavage was performed on each mouse by cannulation of the trachea and lavage with 1 ml of PBS. Cellular content of the bronchoalveolar lavage (BAL) was examined on cytospin preparations stained with Diff-Quick (Dade Behring, Newark, DE) and differential counts were performed on a minimum of 350 cells based on cell morphology and staining. Total cellular counts in the BAL were routinely increased after OVA priming (data not shown). When indicated, statistical significance of eosinophil infiltrate between more than one transfer group was determined using ANOVA. Statistical comparison between two experimental groups was determined using the Student two-tailed t test.

Lungs were prepared for histology by perfusing the right ventricle with PBS. Samples were fixed immediately in 10% formalin for 2 h at room temperature, then stored in 70% ethanol. Tissues were further dehydrated in graded ethyl alcohols and embedded in paraffin. Following sectioning, slides were deparaffinized and stained with H&E or periodic acid Schiff (PAS). For immunohistochemistry, slides were immersed in methanol containing 0.3% H₂O₂ for 30 min to exhaust endogenous peroxidase activity. Sections were stained with a 1/100 dilution of rat-anti-CD3 (Serotec, Raleigh, NC), followed by incubation with a 1/200 dilution biotinylated anti-rat mouse absorbed Ab (Vector Laboratories, Burlingame, CA), and ABC Elite complex (Vector Laboratories), then developed with diaminobenzidine chromogen.

Histology sections were evaluated by C.H., an experienced pathologist who did not know the identity of the experimental groups. An estimate of the qualitative and quantitative degree of cellular infiltration and degree of mucus production and numbers of PAS⁺ cells was made using a scale ranging from 0 (no indication) to 3 (maximum pathology). The location of cellular infiltrate (peribronchial vs perivenular) was also noted. Scoring for inflammation in the H&E sections was as follows: 0, no signs of inflammation; 1, light and dispersed infiltrate in only a few areas of the section; 2, moderate infiltrate surrounding <50% of vessels and airways; 3, heavy and focused infiltrate surrounding the majority of vessels and airways. Scoring for mucus production in the PAS sections was as follows: 0, no signs of mucus in airways or elevated numbers of PAS⁺ cells; 1, no mucus in airways, slight increase in numbers of PAS⁺ cells in a few airways; 2, some mucus detectable in airways, ∼50% of airway epithelial cells are PAS⁺ in multiple airways; 3, mucus plugging of airways observed in several airways, majority of airway epithelial cells are PAS⁺ in multiple airways per section. Scoring of histology sections was performed on four to eight animals per transfer group per experiment and expressed as an average. Areas of the slide that were representative for the whole group were photographed and digitally processed using CoolSnap (Roper Scientific, Trenton, NJ). The images shown are representative of each group.

Splenocytes were isolated from the chimeric mice after priming and challenge as described above. The cells were stimulated in vitro with PMA (20 ng/ml) and ionomycin (1 μM) (Sigma-Aldrich) in the presence of IL-2 (20 U/ml) for 18 h. Production of IL-4 was detected in the supernatants using an ELISA kit from Endogen (Woburn, MA).

Splenocytes were isolated from the chimeric mice after priming twice with OVA as described above. Single cell suspensions of resident lung cells were prepared using a modified method based on Constant et al. (27) and Hogan et al. (28). Lungs were isolated from mice after cannulation of the trachea and perfusion of the airways after priming and aerosol challenge with OVA as described above. Individual lungs were finely minced, resuspended in digestion media (RPMI 1640, 10% FBS, 10 mM HEPES, 1% collagenase (Roche Applied Science, Indianapolis, IN)), and 20 μg/ml DNase I type IV from bovine pancreas (Roche Applied Science), and incubated at 37°C for 1 h on a rotating platform. After 1 h, cells were passed through a 70 μM cell strainer, washed two times with RPMI 1640, and resuspended in PBS containing 5% FCS and 5 mM sodium azide. To reduce nonspecific binding, 1 × 10⁶ cells were incubated with 2.4G2 for 15 min before addition of fluorochrome-labeled mAbs. Monoclonals used were as follows: IAd-PE (clone AMS-32.1; BD PharMingen), CD11b-FITC (clone M1/70), CD11c-FITC (clone HL3), c-kit-PE (clone 2B8), IgE followed by anti-IgE-FITC (clone R35-72), and IL-4Rα-Alexa (M1, a generous gift of Dr. F. Finkelman, University of Cincinnati, Cincinnati, OH). Propidium iodide (Sigma-Aldrich) was added for exclusion of dead cells. Appropriate gates and compensation levels were initially set using single color stains. Data was analyzed using CellQuest (BD Immunocytometry Systems, San Jose, CA). In some cases, FlowJo software (Treestar, Costa Mesa, CA) was used to adjust compensation and calculate percentages.

To delineate the role the IL-4Rα plays in allergic airway inflammation, we induced allergic inflammation in several BM chimeras. Using a comprehensive strategy, chimeric mice were generated using numerous combinations of BM donors and recipients. The chimeric mice analyzed and the expected patterns of expression of the IL-4Rα are described in Table I. Chimeric mice were immunized with Alum or OVA/Alum and nebulized three times with 1% OVA in PBS. We initially analyzed the composition of cells in the BAL (Fig. 1). As expected, OVA priming increased the total number of cells in the BAL compared with Alum alone (data not shown). Chimeric mice expressing IL-4Rα on both BM-derived and non-BM-derived cells (BALB/c BM into RAG2KO recipients) demonstrated a dramatic increase in the percentage of eosinophils in the BAL in response to priming with OVA (5% eosinophils in Alum-treated mice compared with 50% in OVA/Alum-treated mice) (Fig. 1, Panel 1). The percentage of mononuclear cells was reduced proportionally, while the percentage of polymorphonuclear cells and lymphocytes remained largely unchanged. If the recipient mice lacked the IL-4Rα (BALB/c BM into IL-4Rα × RAG2KO recipients), we observed 25% eosinophils in the BAL of OVA/Alum-primed mice compared with the 50% seen in the previous transfer group; however, the percentage of eosinophils was still significantly elevated as compared with that observed in the Alum alone group (Fig. 1, Panel 2). If the BM donor lacked the IL-4Rα (IL-4RαKO BM into RAG2KO recipients), only 5% eosinophils were present in the BAL of OVA/Alum-treated mice (Fig. 1, Panel 3). As expected, we detected a similar increase in total IgE and the presence of OVA-specific IgG1 in the serum of all BM-chimeric mice that expressed the IL-4Rα on both B and T cells and were primed with OVA/Alum (data not shown).

Lung sections from the same transfer groups described in Fig. 1 were prepared and stained with H&E. No cellular infiltrate was present in any of the Alum-primed mice (H&E score = 0) (Fig. 2, a–c). The most severe degree of cellular infiltrate was observed when BALB/c BM was transferred into RAG2KO recipients (H&E score = 3); these chimeric mice are essentially wild type for IL-4Rα expression (Fig. 2, d and g, and Table I). In these chimeric mice, cellular infiltrate was mainly eosinophilic (Fig. 2,g) and was located peribronchially and perivascularly (Fig. 2,d). Interestingly, when the recipient mice lacked the IL-4Rα, eosinophilic infiltrate was still observed, albeit at reduced levels (H&E score = 1.5) (Fig. 2, e and h). If the BM-donor did not express the IL-4Rα (Fig. 2, f and i), the identity and localization of the cellular infiltrate was altered. When IL-4RαKO BM was transferred into RAG2KO recipient mice, the infiltrate was mainly monocytic (Fig. 2,i) and was localized around the veins (Fig. 2,f). The variations seen in the severity and composition of infiltrate correlates to the results seen in the cellular composition of the BAL (Fig. 1). These results indicate that expression of IL-4Rα on non-BM-derived cells is not absolutely required for the recruitment of eosinophils into the lung tissue and airway, but that it does contribute to the degree of inflammation. In contrast, when BM-derived cells, including competent lymphocytes, lack the IL-4Rα, the tissue infiltrate is monocytic instead of eosinophilic. This may be due either to the absence of TH2 cytokines produced in these mice or the unopposed presence of TH1 cytokines.

To visualize goblet cells and mucus, PAS staining was performed on lung sections derived from the same mice (Fig. 3). No PAS⁺ cells were present in the bronchi of Alum-primed mice (data not shown). The most severe degree of goblet cell metaplasia (PAS score = 3) was seen when the IL-4Rα was expressed on both BM-derived and non-BM-derived cells (Fig. 3,a). In addition, these wild-type chimeras had signs of arteriole constriction most visible in these PAS-stained sections (arrows). No PAS⁺ cells were present in recipient mice lacking the IL-4Rα, even though these mice contained IL-4Rα⁺ BM-derived cells (PAS score = 0) (Fig. 3,b). In the absence of the IL-4Rα on BM-derived cells, there was some goblet cell differentiation in the absence of eosinophilic inflammation, but the number of mucus-producing cells was decreased (PAS score = 1) (Fig. 3 c). These results are consistent with those reported by Grunig et al. (5) and Cohn et al. (9) using IL-4RαKO mice. Furthermore, these results clearly establish that even in the presence of BM cell types expressing the IL-4Rα, goblet cell metaplasia and mucus production are dependent upon expression of the IL-4Rα on non-BM-derived cell types, most likely the lung epithelial cells themselves (29).

Although IL-4Rα-deficient mice do not mount a TH2 response, they do have competent lymphocytes that can respond to Ag by producing TH1 cells capable of secreting IFN-γ. When IL-4Rα⁺ mice were reconstituted with IL-4RαKO BM, we found a decrease in the level of cellular infiltrate present in the lungs compared with the wild-type chimera and the infiltrate was mainly monocytic. The presence of a monocytic infiltrate is consistent with the production of TH1 cytokines (4, 30). To avoid complications arising from the production of TH1 cytokines in IL-4RαKO mice, in the remaining experiments we used RAG2KO or IL-4Rα × RAG2KO mice as BM-donors and BM recipients.

It has been well documented that in the absence of the IL-4Rα, precursor T cells do not efficiently differentiate to TH2 cells in response to protein Ags (22). Rather, they mount a predominantly TH1-type response. In addition, numerous studies have shown that the TH2 cytokines are essential to the pathogenesis of allergic asthma. Studies using mice deficient in IL-4 (3), the IL-4Rα (5, 30, 31), or IL-13 (32) have shown that AHR, eosinophilic inflammation, and mucus production were not induced in response to antigenic challenge. To obviate the TH2 deficiency in mice lacking IL-4Rα on the BM, and to define the function of IL-4Rα on non-T cells, OVA-specific TH2 cells prepared in vitro (see Materials and Methods) were transferred into chimeric mice via tail vein injection.

In addition to the wild-type BALB/c, we performed transfers using BM from mice that lacked both the IL-4Rα and lymphocytes (Fig. 4). To directly compare groups that contained wild-type BALB/c BM only to those receiving TH2 cells, OVA priming was performed in all groups. Due to the lack of B cells in the BM transplant, no IgE or OVA-specific IgG1 was detected in the serum of OVA-treated animals (data not shown). In the absence of exogenous TH2 cells, no eosinophils were present in the BAL of RAG2KO mice that received IL-4Rα × RAG2KO BM whether or not they were primed with OVA (Fig. 4,a). This is as expected as no T cells were present in BM derived from RAG2-deficient animals. In the presence of TH2 cells, we found 37% eosinophils in the BAL of these chimeric mice after challenge with inhaled OVA in the absence of priming. In chimeric mice that were primed with OVA, the level of eosinophilia was increased to 65% (Fig. 4 a). Notably, in chimeric mice receiving RAG2KO BM, we found 58% eosinophils in the BAL without priming and 90% eosinophils present after OVA priming. Although the difference between the groups did not achieve statistical significance (p values were <0.08), there was a clear trend. These results indicate that expression of the IL-4Rα on BM-derived cells can contribute to the severity of eosinophilic recruitment in the airway.

These transfers were performed using RAG2KO recipients that expressed the IL-4Rα. Expression of the IL-4Rα on non-BM-derived cells such as the lung epithelium and endothelium is thought to participate in recruitment of eosinophils by signaling the production of eotaxin in response to IL-4 or IL-13 (5, 33, 34). To test the contribution of the IL-4Rα on BM-derived cells in a lung environment that lacks the IL-4Rα, we reconstituted IL-4Rα × RAG2KO recipient mice with the indicated BM in the presence or absence of TH2 cells (Fig. 4, b and c). Two graphs are shown representing the percentage of eosinophils in the BAL from two separate experiments. None of the Alum-primed mice had any eosinophils detectable in the BAL (Fig. 4,b). When BALB/c BM was transferred into IL-4Rα × RAG2KO recipient mice, we found ∼20% eosinophils in the BAL after OVA priming and challenge compared with 10% in mice receiving only TH2 cells (Fig. 4,b). No eosinophils were found in the BAL of mice receiving RAG2KO BM alone (Fig. 4,b). Yet strikingly, there was a substantial increase in the percentage of eosinophils (43%) when mice received both TH2 cells and RAG2KO BM (Fig. 4,b). In the second experiment (Fig. 4 c), IL-4Rα × RAG2KO recipient mice received both TH2 cells and an IL-4Rα × RAG2KO BM supplement. This transfer group demonstrated very low percentages of eosinophils in the BAL in the presence or absence of OVA priming. However, when RAG2KO BM was transferred in the presence of TH2 cells, we detected a significant increase in eosinophils in the BAL after OVA priming (38% eosinophils). These results indicate that expression of the IL-4Rα on a nonlymphocyte, BM-derived cell type(s) contributes to the level of eosinophilic infiltration into the airways. This contribution is most clearly observed in an IL-4Rα-deficient environment.

Lung sections were derived from the animals in the same transfer groups described above and stained with H&E and PAS (Fig. 5). When IL-4Rα × RAG2KO BM was transferred into RAG2KO recipients, we found no signs of eosinophilic infiltrate in the lungs (H&E score = 0). (Fig. 5,a). However, in the presence of TH2 cells, there was an increase in peribronchial and perivascular infiltrate present in the lungs of OVA/Alum-primed mice (Fig. 5,b). The infiltrate consisted predominantly of eosinophils (H&E score = 2.5). Interestingly, transfer of TH2 cells plus RAG2KO BM cells, which express the IL-4Rα, into RAG2KO recipient mice, resulted in a greater degree of eosinophilia in the lung tissue (H&E score = 3) (Fig. 5,c). In addition, in the mice that received TH2 cells and RAG2KO BM, there was an increase in collagen surrounding the bronchi and giant cell granulomas were observed throughout the lung. The effects on PAS staining were similar to the overall differences seen in the H&E-stained sections. We found no PAS-positive cells in RAG2KO recipient mice receiving IL-4Rα × RAG2KO BM alone (PAS score = 0) (Fig. 5,d). In the presence of TH2 cells, we did detect PAS⁺ cells in the airway (PAS score = 1.5) (Fig. 5,e). However, the most intense PAS staining was observed when RAG2KO mice received both TH2 cells and BM from RAG2KO mice (PAS score = 3) (Fig. 5 f). Lung sections were stained for anti-CD3 to identify T cells; T cells were detected in the lungs of chimeric mice that received TH2 cells (data not shown). These results illustrate the requirement of TH2 cells for eosinophilic inflammation, goblet cell metaplasia, and mucus hypersecretion. Furthermore, they suggest that while the expression of the IL-4Rα on BM-derived cells is not absolutely required in an IL-4Rα⁺ environment in the presence of TH2 cells, it does contribute to the severity of inflammation and goblet cell metaplasia.

The contribution of the BM phenotype on the severity of inflammation was more profound in an IL-4Rα-deficient environment (Fig. 6). When BALB/c BM was transferred into IL-4Rα × RAG2KO recipients, low levels of peribronchial and perivenular infiltrate were detected (Fig. 6,a) consistent with results shown in Fig. 2. Lung sections from IL-4Rα × RAG2KO recipient mice that received RAG2KO BM (Fig. 6,b) or TH2 cells alone (Fig. 6,c) demonstrated low or minor cellular infiltrate in the lungs, respectively (H&E scores = 0.5), which correlates with the level of eosinophils detected in the airways (Fig. 4, b and c). However, when TH2 cells and RAG2KO BM were transferred into IL-4Rα × RAG2KO recipient mice (Fig. 6,d), we observed marked elevation of eosinophilic inflammation that was located peribronchially and perivascularly (H&E score = 3). Giant cell granulomas were also observed. Of note, none of these sections stained positive for PAS (Fig. 6, a–d, insets).

To determine whether the transferred TH2 cells were able to localize to the lung, immunostaining with anti-CD3 was performed on lung sections (Fig. 6, e–h). The presence of T cells in the lungs of IL-4Rα × RAG2KO recipient mice that received BALB/c BM (Fig. 6,e), TH2 cells alone (Fig. 6,g), or TH2 cells and RAG2KO BM (Fig. 6,h) was confirmed by immunohistochemical staining. No staining was apparent in lung sections from mice that received RAG2KO BM alone as expected because this transfer group lacks endogenous lymphocytes (Fig. 6,f). In addition, we tested the levels of IL-4 production by activated splenocytes to confirm that functional TH2 cells were still present in the chimeric mice (Fig. 7). Increased concentrations of IL-4 were detected in the supernatants of splenocytes isolated from mice that received TH2 cells, but not from mice that did not receive the TH2 cells. Similar results were obtained for IL-13 production (data not shown). These results confirm that the T cells injected into the mice were present in the lung and spleen, and were still able to produce TH2 cytokines. Because there were TH2 cells and substantial eosinophilic inflammation observed in the lungs from IL-4Rα × RAG2 KO mice that had received TH2 cells and RAG2KO BM (Fig. 6, d and h), these results emphasize the requirement for expression of the IL-4Rα on non-BM-derived cells for the development of goblet cell metaplasia and mucus hypersecretion. Importantly, they indicate that expression of the IL-4Rα on BM-derived cells contributes substantially to the eosinophilic infiltrate mediated by TH2 cells.

These data indicate that eosinophilic inflammation in the lung is regulated by a nonlymphoid, BM-derived cell whose functional contribution requires expression of the IL-4Rα and Ag stimulation. To begin to identify this cell type(s), we analyzed the phenotype of cells in the lung after immunization with OVA and aerosol challenge and in the spleen after stimulation with OVA (Figs. 8 and 9). We focused on physical characteristics (forward scatter and side scatter) and surface Ags characteristic of mast cells, dendritic cells, and monocytes as described in Materials and Methods. The candidate effector cell should express the IL-4Rα, should be elevated in mice receiving RAG2KO BM, as compared with IL-4Rα × RAG2KO BM, and should be increased by OVA immunization. We first analyzed cells in the lung after OVA priming and challenge (Fig. 8). We analyzed cytospins of the lung cells isolated from the individual animals. Based on differential staining, we found preferential expansion of eosinophils in the IL-4Rα⁺ BM recipients (Fig. 8,a) as expected. No other granulocytes were observed in the population. Using FACS analysis on pooled lung cells, we found a preferential expansion of cells with high side scatter characteristics in the animals that received IL-4Rα⁺ BM (40%) as compared with those that received Il-4Rα-deficient BM (25%); a marked proportion of these cells (30% of high scatter, 14% of total) expressed the IL-4Rα chain (Fig. 8,b). Interestingly, in the animals that received IL-4Rα⁺ BM, the percentage of CD11c⁺, IAd⁺ of an intermediate intensity (IAd^int, mean fluorescence intensity (MFI) 100) cells were increased by priming with OVA (4.2% Alum vs 9.0% OVA/Alum), but not in the animals that received IL-4Rα-deficient BM (4.7% Alum vs 3.6% OVA/Alum) (Fig. 8,c). These CD11c⁺, IAd^int cells were also IL-4Rα⁺ (Fig. 8 c) and were in the high side scatter region (data not shown). The percentage of cells expressing the IL-4Rα⁺ was similar to the percentage of CD11c⁺, IAd^int (4% Alum vs 9.0% OVA/Alum). The high side scatter cells were also IAd^int (MFI 100), CD11b⁺ (data not shown). Taken together, these results suggest that the cell type most affected by OVA priming in the lung is IL-4Rα⁺, high side scatter, IAd^int, CD11c⁺, CD11b⁺. It is possible these cells represent a subset of the infiltrating eosinophils which have been reported to express MHC class II, CD11b, and CD11c under certain conditions (35, 36, 37, 38).

To characterize cells that might preferentially expand during Ag stimulation, we also analyzed cell surface phenotypes of splenocytes. Strikingly, we found a 30-fold increase in the surface expression of IAd on spleen cells if the BM was derived from an IL-4Rα⁺ donor (MFI 300) as compared with an IL-4Rα-deficient donor (MFI 9) in the presence of TH2 cells (Fig. 9,a). In the animals that received IL-4Rα⁺ BM, the percentage of IAd⁺ cells was greatly amplified by priming with OVA (14.2% Alum vs 51.3% OVA/Alum), but not in the animals receiving IL-4Rα-deficient BM (5.1% Alum vs 9.5% OVA/Alum). The majority of the IAd⁺ cells were also CD11b⁺ (Fig. 9,b). We did not detect substantial differences in the percentages of CD11c⁺, IAd⁺ cells (all ∼4%). The IAd^high cells were of low side scatter (data not shown) and demonstrated detectable levels of IL-4Rα (Fig. 9 a). These results suggest that an IL-4Rα⁺ monocytic cell in the spleen may contribute to allergic inflammation.

The studies reported herein elucidate several roles for the IL-4Rα in regulating the evolution of hallmark pathologies associated with allergic airway inflammation. Although previous reports indicated the general importance of IL-4, IL-13, and the IL-4Rα (5, 30, 31), we show that expression of the IL-4Rα on both BM-derived cells and non-BM-derived cells contributes to disease progression and the nature of the cell type expressing the IL-4Rα determines the prevailing pathological characteristics of the disease.

We found that the IL-4Rα expressed on both BM cells, which are not lymphocytes, and non-BM cells participates in the regulation of eosinophil infiltration into the lung and airways. Furthermore, these studies clearly support the critical importance of IL-4Rα expression on non-BM cells for goblet cell hyperplasia and mucus hypersecretion. The segregation of biological responses mediated by the IL-4Rα seen in our studies is consistent with results observed using infectious models. Interestingly, experiments using intestinal helminth parasites also indicate a unique role for the IL-4Rα on BM-derived cells vs non-BM-derived cells in regulating anti-helminth responses in the gut (23).

It is now well established that TH2 cytokines are critical for the development of asthma. In the absence of IL-4, the IL-4Rα, or STAT6, T cells do not efficiently differentiate to TH2 cells (22, 39, 40, 41, 42) or develop the hallmark symptoms of allergic inflammation. Both IL-4 and IL-13 signal through the same receptor subunit, IL-4Rα, in a STAT6-dependent manner. This sharing of the IL-4Rα results in many biological functions that can be regulated by either cytokine, including mucus production and eosinophilic inflammation. However, some reports indicate IL-13 is the principal cytokine for mucus hypersecretion and AHR in vivo (5, 6, 32), while IL-4 is critical for T cell development and the recruitment of cells to the lungs (4). Regardless of the relative contribution of these two ligands, our findings illustrate the importance of IL-4Rα expression on specific cells in directing the course of the disease.

Numerous studies have determined that the key pathological characteristics of allergic lung inflammation including AHR, mucus production, and cellular inflammation are not dependent upon one another (9, 30, 32, 43). Our results are consistent with this notion. For example, in some transfer groups we were able to detect massive eosinophilic inflammation in the absence of mucus hypersecretion and goblet cell hyperplasia. Additionally, in mice demonstrating profound mucus production and eosinophilic inflammation, we were unable to detect induction of AHR in the mice using whole body plethysmography (data not shown).

Increased levels of IgE and OVA-specific IgG1 were detected in the serum of all BM-chimeric mice immunized with OVA-Alum and expressing the IL-4Rα on both B and T cells (data not shown). Several studies have shown the while IgE is required for many type I allergic responses, neither IgE, nor any other Abs are required for the pathogenesis of allergic inflammation in mice (3, 44, 45). In support of this finding, we observed substantial inflammation and mucus production in chimeric mice on the RAG2-deficient background, which lack lymphocytes, in the presence of TH2 cells without induction of IgE or Ag-specific IgG1.

Our results combined with previous work clearly demonstrate that expression of the IL-4Rα on non-BM-derived cells is required for mucus production. In the absence of the IL-4Rα on non-BM-derived cells, we found no evidence of mucus production or goblet cell hyperplasia, even in the presence of TH2 cells and eosinophils. The lung epithelial cells are most likely the cells that require expression of the IL-4Rα to regulate mucus production (29, 43). Direct addition of IL-4 or IL-13 by transgenic targeting (46, 47) or by intranasal instillation (5) can induce the production of mucus in RAG2KO mice. Furthermore, these cytokines have been shown to induce expression of Muc5 AC mRNA in epithelial cells (48). Therefore, the induction of goblet cell metaplasia and mucus hypersecretion appears to be directly regulated by signaling in epithelial cells via the IL-4Rα by IL-4 and/or IL-13.

As mentioned earlier, while IL-4Rα-deficient mice do not mount a TH2 response, they do have competent lymphocytes that can respond to Ag by producing TH1 cells capable of secreting IFN-γ. When IL-4Rα⁺ mice were reconstituted with IL-4RαKO BM, we found a decrease in the level of cellular infiltrate present in the lungs compared with the wild-type chimera and the infiltrate was mainly monocytic. The presence of a monocytic infiltrate is consistent with the production of TH1 cytokines (4, 31). To avoid complications arising from the production of TH1 cytokines in IL-4RαKO mice, we used RAG2KO or IL-4Rα × RAG2KO mice as BM donors and BM recipients in the presence or absence of OVA-specific TH2 cells.

Our results indicate that while overall expression of the IL-4Rα is required to mediate the entry of eosinophils into the lungs, expression on both BM-derived and non-BM-derived cell types is not absolutely required. The expression of the IL-4Rα on one of these compartments is sufficient to support eosinophilic inflammation, but expression on both contributes to the severity of infiltration and the degree of goblet cell hyperplasia. In patients that die of severe acute asthma, the hallmark feature in the lungs is the presence of excessive mucus resulting in almost complete occlusion of the airways (49). Therefore, contribution to the regulation of severity of allergic lung inflammation and mucus hypersecretion may have significant physiological relevance.

Based on the comparisons between IL-4Rα-positive and -negative recipients, it would appear that expression on a cell of non-BM origin is dominant in regulating the level of inflammation. Alternatively, it is possible that radiation-resistant BM-derived cells, including monocytes and dendritic cells, remaining in the IL-4Rα⁺ recipient mice participate in the eosinophilic inflammation.

The degree of eosinophilic inflammation is determined by several factors including mobilization of cells from the BM by IL-5, the adhesion of cells to the vascular endothelium of the lung, and the recruitment of eosinophils by chemokines (33, 34). It has been clearly established that IL-4 induces expression of VCAM-1 on endothelial cells and that eosinophils interact with this adhesion molecule when migrating into the lung (50). However, T cells and eosinophils were capable of entering the lungs of mice that did not express the IL-4Rα. We were unable to detect any differences in VCAM-1 expression on the lungs of the IL-4Rα⁺ or IL-4Rα⁻ recipient mice (data not shown), suggesting that entry of eosinophils and T cells into the lung tissue do not require IL-4Rα expression in recipient mice under the conditions used in these studies.

There are a number of chemokines involved in the allergic inflammatory response. Several chemokines, including eotaxin, recruit eosinophil migration to the lungs. Interestingly, the epithelial cells, endothelial cells, fibroblasts, and smooth muscle cells of the lung are all capable of producing chemokines in response to a variety of stimuli and several have been shown to produce eotaxin in response to IL-4 and IL-13 (34). Based on our studies, it appears that the IL-4- or IL-13-induced chemokine production by these cell types is not required to support eosinophil recruitment. Furthermore, by quantitative RT-PCR for a number of chemokines including, eotaxin, RANTES, monocyte chemotactic protein-3, macrophage inflammatory proteins 1α and 1β, we found that OVA priming generally resulted in an increase in chemokine mRNA in the lungs of IL-4Rα × RAG2KO mice, whether they were recipients of either IL-4Rα⁺ or IL-4Rα-deficient BM (data not shown). Thus, we cannot explain the differences in levels of inflammation in these chimeras to differences in chemokine production. It is possible the difference may be in the ability of cells to respond appropriately to the chemokines.

It is possible that because we transferred complete BM into our recipient mice, an epithelial stem cell could repopulate the epithelium with IL-4Rα⁺ cells in the lungs of IL-4Rα-deficient mice. If this did occur, then the IL-4Rα⁺ epithelial cells could provide chemokines that would be able to recruit eosinophils to the lungs. Although we cannot directly rule this out, this explanation is unlikely because we found no evidence of goblet cell development in IL-4Rα-deficient recipients of IL-4Rα⁺ BM and TH2 cells or differences in chemokine production.

The addition of both TH2 cells and IL-4Rα⁺ BM cells to IL-4Rα-deficient mice supported eosinophilic inflammation into the lungs after priming and aerosol challenge with OVA. These data indicate that eosinophilic inflammation is regulated by expression of the IL-4Rα on a BM-derived cell that is not of lymphoid origin that requires Ag stimulation. This cell type may be involved in amplifying the TH2 response and/or may develop effector functions in response to signaling to IL-4 or IL-13. There are a number of different cells that could fall into this category. One such cell type is the dendritic cell. A number of studies have recently been published indicating that BM-derived dendritic cells are critical in the differentiation and activation of TH2 cells in the lung (51, 52). However, by FACS analysis of spleen and lung cells, we did not observe a substantial change in percentage of CD11c⁺, IAd^high cells. A second candidate is the mast cell. Several studies indicate that murine mast cells, while not required, do play a role in regulating the severity of AHR and airway inflammation in an IgE-dependent manner (53, 54, 55). Because we used RAG2-deficient mice, the absence of IgE would have prevented activation of mast cells via cross-linkage of the FcεRI. It is interesting to note that IL-4 regulates the expression of leukotriene C4 synthase in mast cells, which could affect their functional potency (56). However, we did not detect any c-kit⁺, IgER⁺, mast cells in the spleen or lung of treated mice (data not shown).

A third candidate is the eosinophil itself as studies have shown that signaling through the IL-4Rα primes eosinophils for chemotaxis (17). We found a preferential expansion of eosinophils in the lungs by staining and of IL-4Rα⁺ high scatter cells by FACS in mice receiving RAG2KO BM in response to OVA immunization and challenge. Interestingly, the IL-4Rα⁺ high scatter cells were also IAd^int, CD11b⁺, CD11c⁺. Because the numbers of eosinophils were enhanced in the lung and lung cells of high side scatter in allergic inflammation have been shown to be largely comprised of eosinophils (37), we suspect these cells may represent a subset of eosinophils which have been reported to express MHC class II, CD11b, CD11c and possess some ability to present Ag to T cells (35, 36, 37, 38).

A fourth candidate is the monocyte. IL-4 and IL-13 have many effects on monocytic cells that may contribute to inflammation including the induction of MHC class II, vasoactive compounds, and chemokines (33). Indeed, in the spleen we observed preferential expansion after OVA priming of CD11b⁺, IAd⁺, IL-4Rα⁺, low scatter cells in animals receiving IL-4Rα⁺ BM. Stimulation of such cells by IL-4 or IL-13 may help to establish an environment that serves to amplify the TH2-mediated inflammatory response, perhaps through their ability to stimulate the up-regulation of MHC class II, and/or stimulate participation in the evolution of the inflammatory response. In this regard, it has recently been shown that an IL-4Rα⁺, CD11b⁺ monocyte, that requires IL-4 stimulation, participates in the increase in vascular permeability seen in a mouse model of anaphylaxis, possibly via the production of platelet-activating factor (57, 58).

In summary, we show that the IL-4Rα plays unique and complex roles in a murine model of allergic inflammation depending upon whether the receptor is expressed on BM-derived cells or non-BM-derived cells. We propose that signaling via the IL-4Rα on BM-derived cells, such as eosinophils and monocytes, contributes to the degree of lung inflammation while signaling on lung epithelium regulates mucus production (Fig. 10). The various cellular responses to IL-4 and/or IL-13 likely act in concert to establish a permissive environment and to elicit the full spectrum of pathologies found in asthma. Future studies will be aimed at fully characterizing the BM-derived, MHC class II⁺, nonlymphoid cell types and analyzing the mechanism by which they regulate allergic inflammation. The targeting of such cells could be helpful in the development of novel therapeutic strategies for asthma.

We thank Drs. David Corry, Gabriele Grunig, and Andrea Keane-Myers for helpful discussions and insight during the course of this work. We thank Dr. Fred Finkelman for providing M1-Alexa and Dr. John Ryan for providing the protocol and reagents for mast cell staining. We also thank Drs. Mark Williams and Alice Welch for advice on FACS analysis. We acknowledge Elena Semanova and Xiulan Qi for excellent technical assistance.