Prolonged Eosinophil Production after Allergen Exposure in IFN-γR KO Mice is IL-5 Dependent


  • L.-L. Zhao,

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author
  • J. Lötvall,

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author
  • A. Lindén,

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author
  • M. Tomaki,

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author
  • M. Sjöstrand,

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author
  • A. Bossios

    1. Lung Pharmacology Group, Department of Internal Medicine/Respiratory and Allergology, The Sahlgrenska Academy, Göteborg University, Gothenburg, Sweden
    Search for more papers by this author

Apostolos Bossios, MD, PhD, Lung Pharmacology Group, Department of Internal Medicine/Respiratory Medicine and Allergology, The Sahlgrenska Academy, Göteborg University, Guldhedsgatan 10A, SE-413 45 Gothenburg, Sweden. E-mail:


Asthma is a T helper 2 (Th2)-driven inflammatory process characterized by eosinophilia. Prolonged airway eosinophilia is commonly observed in asthma exacerbations. Our aim was to evaluate whether eosinophilia in prolonged allergic inflammation is associated with a continuous supply of new eosinophils to the airways, and how this is regulated. Ovalbumin (OVA)-sensitized interferon-γ receptor knockout mice (IFN-γR KO), known to maintain a long-lasting eosinophilia after allergen exposure, were compared to wild type (wt) controls. Animals were exposed to OVA or phosphate-buffered saline on three consecutive days, and bone marrow (BM), blood and bronchoalveolar lavage (BAL) samples were collected 24 h, 7 and 21 days later. Newly produced cells were labelled using bromodeoxyuridine (BrdU). Serum IL-5 was measured and its role was investigated by administration of a neutralizing anti-IL-5 antibody. In-vitro eosinophilopoiesis was examined in both groups by a colony-forming assay. Allergen challenge increased eosinophils in BM, blood and BAL, in both IFN-γR KO and wt mice, both 24 h and 7 days after the last allergen exposure. At 21 days after the last exposure, only IFN-γR KO mice maintained significantly increased eosinophil numbers. Approximately 50% of BAL granulocytes in IFN-γR KO were produced during the last 6 days. Interleukin (IL)-5 concentration was increased in IFN-γR KO mice, and anti-IL-5 reduced eosinophil numbers in all compartments. Increased numbers of eosinophil colonies were observed in IFN-γR KO mice after allergen exposure versus controls. In this model of a Th2-driven prolonged allergic eosinophilia, new eosinophils contribute to the extended inflammation in the airways by enhanced BM eosinophilopoiesis in an IL-5-dependent manner.


Allergic asthma is a chronic inflammatory disease of the airways characterized by cell infiltration after allergen exposure. There is a recruitment of a variety of inflammatory cells, i.e. CD4+ T lymphocytes, basophils, neutrophils, macrophages and eosinophils, which interact with a variety of structural cells like epithelial and smooth muscle cells. Result of above is a chronic inflammation of the bronchial mucosa. Among them eosinophils represent a major inflammatory cell contributing in many clinical features of asthma as variable airway obstruction and bronchial hyperresponsiveness. Eosinophils provoke through production and release toxic products as lipid mediators and basic granule proteins, as well as several cytokines [1–3]. Recently, the role of eosinophils in asthma was questioned after failure of an early clinical trial with anti-IL-5 [4], however, newer studies have re-established their putative involvement in airway wall remodelling [5, 6]. Many asthmatic patients have a prolonged eosinophilic airway inflammation which is further enhanced during periods of asthma exacerbations, as those mimed by allergen exposure [7–10].

Although similar pathological features can be found in both allergic and non-allergic asthma, and the fundamental mechanisms underlying the development of the asthmatic state remain largely elusive, a dominance of a T helper 2 (Th2) over a T helper 1 (Th1) phenotype is considered as important for asthma development and orchestration [11–13]. The archetype cytokines that characterize Th1/Th2 are interferon-γ (IFN-γ)/interleukin (IL)-4. IL-4 plays a pivotal role in allergic diseases as cause immunoglobulin E (IgE) switching in B cells; enhance IgE production and therefore atopy [2]. Even more, IL-4 can regulate the polarization of immune response into Th1 or Th2 [14]. Importantly, IL-4 also facilitate the transit of eosinophils across endothelial and epithelial barriers by upregulation on vascular endothelium expression of vascular cell adhension molecule (VCAM)-1, the specific ligand for eosinophil-specific adhension molecule, integrin β1; very late antigen (VLA)-4 [15]. Of other Th2 cytokines, IL-5 plays a crucial role for the induction of eosinophilia as being the most potent survival factor and a main proliferative cytokine for their precursors [16–18]. However, recent data implicate a specific role of IL-5 in mobilization of eosinophils [19] as well in their release from the bone marrow (BM) and their maturation from progenitor cells [20].

Increased evidence suggests that BM eosinophilopoiesis is enhanced in allergic patients as well as in animal models of allergen-induced inflammation [21, 22]. For example, in atopic asthmatic subjects, the levels of CD34+ progenitors cells and IL-5 responsive Eo/Baso- colony-forming units (CFUs) are raised in both peripheral blood (PB) and BM [23, 24]. Furthermore, after allergen exposure, there is a significant increase in the BM CD34+ cells which express IL-5Rα, which is paralleled with a significant increased sputum eosinophilia and increased airway responsiveness [25]. However, it is not clear whether a prolonged eosinophilia in the airways depends fully on prolonged survival of eosinophils entering the site of inflammation at an early phase, or whether there is a continuous traffic of new cells from the BM to the airways.

The aim of this study was therefore to determine whether newly produced eosinophils are recruited continuously to the airways, based on a hypothesis that chronic airway inflammation is a dynamic process dependent at least partly on prolonged enhanced eosinophilopoiesis. To evaluate this, we use a model of an enhanced Th2 phenotype; which is mice lacking the receptor for IFN-γ knockout mice, (IFN-γR KO mice). These mice respond to airway allergen exposure with prolonged airway eosinophilia being maintained weeks after the end of allergen exposure [26]. Two weeks after end of allergen exposure mice were given 5-Bromo-2′-deoxyuridine (BrdU), which incorporates into the DNA during S-phase of the cell cycle, and is utilized as a marker of cells that have undergone mitosis since the BrdU injection. Furthermore, the role of the key Th2 cytokine IL-5 was also investigated by administration of a neutralizing anti-IL-5 antibody. Finally, the ability of CD34+ BM progenitors to generate eosinophils was compared in IFN-γR KO and wild type (wt) mice in-vitro by culture in a semi-solid medium.

Materials and methods

Animals.  Male IFN-γR KO, 6–8 weeks of age, were kindly donated by Professor N. Lycke (Department of Clinical Immunology, University of Göteborg, Sweden). Male 129/Sv/Ev F2 mice 6–8 weeks of age were purchased from B&K Universal AB (Sollentuna, Sweden). All mice were provided with food and water ad libitum and housed in specific pathogen free animal facilities. The Ethical Committee for animal studies at Göteborg University approved this study.

Sensitization and airway challenge protocol.  Mice were sensitized intraperitoneally (i.p.) twice (on day 0 and day 7) using 10 μg chicken egg ovalbumin (OVA) (Sigma Chemical Company, St. Louis, MO, USA) adsorbed to 1 mg aluminium hydroxide [Al (OH)3], suspended in 0.5 ml PBS (phosphate-buffered saline). A common sensitization rout in mouse models of allergic inflammation. Two weeks after the second sensitization, and exposed intranasally to 100 μg of OVA dissolved in 50 μl PBS or plain PBS. Mice were exposed to allergen on three consecutive days. bronchoalveolar lavage (BAL), blood and BM were taken 24 h, 7 and 21 days after the last exposure.

To label newly produced cells during late phase of inflammation, a group of animals were given BrdU (Roche, Diagnostics Scandinavia AB, Bromma, Sweden) at a dose of 1 mg in 250 μl PBS by i.p.(twice, 8 h apart) on days 6 and 3 before last sample harvest.

Additionally for evaluation of IL-5 role in late inflammation, in a separate set of experiments, an anti-mouse IL-5 monoclonal antibody (mAb) (Clone TRFK5, 50 μg/animal; R&D Systems Europe Ltd, Abingdon, UK) or its isotype control (rat IgG1, Clone R3-34; Pharmingen, San Diego, CA, USA) dissolved in 0.3 ml PBS was given i.p (1 h prior to the first BrdU injection) on day 6 before last sample harvest (Fig. 1).

Figure 1.

 The experimental design. Mice were sensitized i.p. twice (on day 0 and day 7), using 10 μg OVA. Two weeks after the second sensitization, the allergen or vehicle was given intranasally. Mice were exposed on three consecutive days. BM, blood and BAL were taken 24 h, 7 or 21 days after the last exposure. Mice were given BrdU twice on each injection day (8 h apart). BrdU was injected in separate groups of mice at the time points 3 and 6 days, prior to the sample taken. Finally in another group, anti-IL-5 Ab; 50 μg/animal was given i.p. 6 days prior samples taken.

Cell collection and sample processing.  The mice were anaesthetized by an i.p. injection of xylazine 5 mg/kg (Bayer Sverige AB, Gothenburg, Sweden) and ketamine 50 mg/kg (Park-Davis Scandinavia, Stockholm, Sweden).When in adequately deep anaesthesia chest was opened and blood was obtained by penetration of the right ventricle of the heart with a needle. BAL was performed by instilling 0.5 ml of PBS through a tracheal cannula, followed by gentle aspiration and repeated with 0.5 ml PBS. Finally, BM was harvested by excising one femur, which was cut at the epiphyses and flushed with 4 ml of PBS. BAL fluid and BM cell suspension was kept on ice until further processing.

Cytospin preparations of blood were obtained by mixing 200 μl with 800 μl of 2 mm EDTA (Sigma-Aldrich) in PBS, followed by red blood cells (RBC) lysis in 0.1% potassium bicarbonate and 0.83% ammonium chloride for 15 min at room temperature (RT). White blood cells (WBC) were resuspended in PBS containing 0.03% bovine serum albumin (BSA; Sigma-Aldrich). For measurement of IL-5 in serum the remaining volume of blood was centrifuged at 800 g for 15 min at 4 °C and samples were stored at −80 °C until analysis.

Bone marrow and BAL fluid samples were centrifuged at 300 g for 10 min at 4 °C. The cells were resuspended with 0.03% BSA in PBS. The total number of cells in the blood, BM and BALF were determined using standard haematological procedures. Cytospins of blood, BM and BALF samples were prepared and stained according to the May–Grünwald–Giemsa method for differential cell counts. Cell differentiation was determined by counting at least 300 cells using a light microscope (Zeiss Axioplan 2; Carl Zeiss, Jena, Germany). The cells were identified using standard morphological criteria [17]. The cytospin preparations for immunocytochemistry were fixed (4% paraformaldehyde in PBS for 20 min), washed (15% sucrose in PBS for 10 min), then air-dried overnight and stored at −80 °C until further examination.

IL-5 assay.  Mouse IL-5 levels in serum were detected using commercial murine IL-5 enzyme-linked immunosorbent assay kit (R&D Systems, Inc, Abingdon, UK) according to the manufacturer’s instruction. The sensitivity of detection was 7.8 pg/ml.

Immunocytochemical detection of BrdU-labelled cells.  5-Bromo-2′-deoxyuridine incorporated into cellular DNA in cytospins of BAL preparations were detected by immunocytochemistry using a rat mAb against BrdU. The paraformaldehyde-fixed cytospin preparations were washed with tris-buffered saline (TBS) and placed in 0.1% trypsin and 0.1% CaCl2 in PBS at 37 °C for 15 min. The slides were further incubated in 4 m HCl for 5 min to denature the DNA, followed by 0.1 m boric acid (pH 8.5) for 10 min. Non-specific binding sites were blocked with 5% normal rabbit serum (Dako A/S, Glostrup, Denmark) at RT for 15 min. Subsequently, the slides were incubated with 2.5 μg/ml rat anti-BrdU Ab (clone BU1/75; Harlan Sera-Lab, Loughborough, UK) or isotype control (rat IgG2a, clone R35-95; Pharmingen) at RT for 1 h. After washing in 0.05% Tween/TBS and TBS, the slides were incubated with rabbit anti-rat Ig Ab (DAKO A/S) for 30 min and then incubated with rat APAAP (DAKO A/S) for 30 min. Fast Red Substrate System solution (DAKO Corporation, Carpinteria, CA, USA) was used and 20 min elapsed until a clearly visible colour developed. The slides were then counterstained for 1 min with Mayer’s haematoxylin.

Bone marrow cell culture.  Twenty-one days after the last allergen exposure, 30 IFN-γR KO and 30 wt control mice were euthanized. In this set of experiments, both femurs of each mouse were removed to generate as many cells as possible. The femurs were opened at the ends and a needle was inserted. The BM cells were collected by flushing with 4 ml of PBS with 0.5% BSA (PBS-BSA). In the different experiments, cells from 9 to 10 animals from each group were pooled.

CD34+ progenitor cells were enriched from the BM using a magnetic cell sorting system (MACS; Miltenyi Biotec GmbH, Gladbach, Germany). BM cells in a single cell suspension, were washed resuspended in PBS-BSA and labelled with a biotinylated mAb directed to CD34+ (clone RAM34; Pharmingen) in a concentration of 0.5 μg/μl in PBS-BSA. After incubating and washing the cells, 10 μl of streptavidin magnetic microbeads/107 cells was added, according to the manufacturer’s instructions (MACS). The magnetically labelled CD34+ cells were enriched on a positive selection column over a magnetic field. The enrichment procedure resulted in approximately 70% purity of CD34+ cells, according to immunocytochemical staining.

The CD34+ cells were seeded in a concentration of 0.25 × 106/ml in 0.9% metylcelluose enriched Iscoves media (IMDM; GibcoBRL, Paisley, UK) supplemented with 20% FCS, 1% Penicillium–Streptomycin, 2 mm l-glutamin and 10 mm mercaptoethanol in 12-well plates. Mouse recombinant IL-5 (2.5 ng/ml) (R&D Systems Ltd, Oxford, UK) was also added to the media. After 8–10 days of culture by incubation in 37 °C, 7% CO2 in a humidified incubator, the number of colonies with more than 50 cells were counted. Three wells per experiment and group were cultured.

Statistical analysis.  Data are expressed as mean ± SEM, or median (p25, p75): Statistical analysis was carried out using a non-parametric analysis of variance (Kruskal–Wallis test) to determine the variance among more than two groups. If significant variance was found, an unpaired two-group test (Mann–Whitney U-test) was used to determine the significant differences between individual groups. < 0.05 was considered statistically significant.


The effect of OVA exposure on eosinophilia, newly produced granulocytes and IL-5 production

Airway allergen exposure caused a significant increase in BM, blood and BAL eosinophils in both sensitized wt control mice and IFN-γR KO mice, and there were no significant differences between these two groups 24 h and 7 days after the last exposure. However, 21 days after the last exposure, IFN-γR KO mice had substantially higher eosinophil levels in BM, blood and BAL as compared to wt mice (Fig. 2A).

Figure 2.

 Allergic Inflammation after allergen exposure in sensitized animals. (A) Time course of BM, blood and BAL eosinophilia in OVA sensitized and exposed IFN-γR KO or wild type mice. The figure also shows the blood and bone marrow eosinophil level of naive IFN-γR deficient and wild type mice. Mann–Whitney U-test showed significant variance among groups 21 days after allergen exposure. (*P < 0.05, **P < 0.01). (B) The number of BAL BrdU-positive granular cells in IFN-γR KO and wild type mice 21 days after the last allergen exposure. BrdU was given by the i.p. route 3 and 6 days prior to BAL. Filled part of the columns illustrates BrdU-positive cells. Open part of the columns illustrates total granulocytes. Mann–Whitney U-test showed significant variance among groups 21 days after allergen exposure. (***P < 0.001 between wt and IFN-γR KO). (C) Serum IL-5 levels in IFN-γR KO and wild type mice 21 days after the last allergen exposure. Mann–Whitney U-test showed significant variance among groups 21 days after allergen exposure, and a significantly increased level of IL-5 is seen in allergen exposed IFN-γR KO mice compared to wt mice (*P < 0.05). Five to eight mice per group were used, with similar results, and data are shown as mean ± SEM.

To investigate the production of new eosinophils, during late inflammation we injected BrdU 6 and 3 days before harvesting cells from the mice. BrdU-positive granulocytes were rarely found in BAL in wt mice but were abundant in IFN-γR KO mice 21 days after allergen exposure, with a relative number of 51% (Fig. 2B). A vast majority of BrdU-positive granulocytes are eosinophils, i.e. newly produced eosinophils, as the average relative number of neutrophils is <4% in any group. Specifically, the analysis of BAL cell differentiation at 21 days after last allergen exposure with May–Grünwald–Giemsa showed that eosinophils were the main cell that was increased in IFN-γR KO mice compared to wt, followed by lymphocytes. Importantly, there was no difference in the number of neutrophils between groups which even more it was low (mean/median: wt 1.9/1.5, IFN-γR KO 3.6/2.5) (Table 1). So we can agree that the vast majority of BrdU-positive granulocytes in BAL are eosinophils. BrdU-stained BAL cells are shown in the photomicrograph (Fig. 3).

Table 1.   Differential count of BAL cells 21 days after allergen exposure.
Mice phenotypeTotal cells (×104)Macrophages %Neutrophils %Eosinophils %Lymphocytes %
  1. Wild type or IFN-γR KO mice were OVA sensitized and exposed. BAL cells were collected 21 days after last exposure. Differential cell count was performed by staining with May–Grünwald–Giemsa. Five to eight mice per group were used, and data are shown as median (p25, p75). There is an increase in total cell number in IFN-γR KO mice compared to wt (*P = 0.001, Mann–Whitney). In wt mice, the majority of cells are macrophages. In IFN-γR KO, we have a decrease in macrophage percentage (*P = 0.001, Mann–Whitney) as a result of the increase of mainly eosinophils (*P = 0.001, Mann–Whitney) and secondly lymphocytes (*P < 0.01, Mann–Whitney). There is no difference in the percentage of neutrophils.

wt4 (2, 15)91.5 * (82.5, 96)1.5 (0.5, 3)1 (0, 6)5.5 (1.5, 7)
IFN-γR KO28.85* (27.1, 37.5)34 (28.6, 44.4)2.5 (1.6, 4.1)47.8 * (42.5, 52.4)11.5 * (8.8, 14)
Figure 3.

 Representative photomicrographs of BAL cells cytocentrifuged preparations stained positively for BrdU with Fast Red Substrate System in IFN-γR deficient mice (A) and wild type mice (B) with original magnification ×1000. All animals were sensitized and exposed to OVA. Positive BrdU staining is red. In IFN-γR KO mice, a substantial number of BrdU-positive granulocytes were found, and these granulocytes are mainly eosinophils, as neutrophils are very few. In wild type mice, some BrdU staining in macrophages could be found.

To confirm and establish the Th2 phenotype that occurs in IFN-γR KO mice, we measured IL-5 in serum, which was shown to be significantly elevated in IFN-γR KO mice as compared to wt mice in samples taken 21 days after the last allergen exposure (Fig. 2C).

The effects of anti-IL-5 on late eosinophilia

Pretreatment with the neutralizing anti-IL-5 antibody significantly decreased the maintained eosinophils in BM, blood and BAL in IFN-γR KO mice 21 days after the last allergen exposure (Fig. 4A), to the levels of wt mice. Furthermore, anti-IL-5 reduced the number of BrdU-positive granulocytes in BAL in the same animals (Fig. 4B).

Figure 4.

 Effects of anti IL-5 on late allergic inflammation. OVA sensitized and exposed IFN-γR KO or wild type mice were sampling 21 days after the last exposure. Anti-IL-5 or isotype IgG1 was given at a dose of 50 μg/animal, 6 days prior samples taken. BrdU was given by the i.p. route 3 and 6 days prior to BAL. (A) BM, blood and BAL eosinophils. Mann–Whitney U-test showed significant variance among groups 21 days after allergen exposure. **P < 0.01, ***P < 0.001 between wt and IFN-γR. (B) The number of BAL BrdU-positive granular cells in IFN-γR KO. Anti-IL-5 significantly decreased the number of BrdU-positive granulocytes in BAL. Six to eight mice per group were used, with similar results and data are shown as mean ± SEM.

Bone marrow eosinophil colonies

Bone marrow CD34+ progenitor cells from OVA sensitized and exposed IFN-γR KO mice cultured in the presence of rmIL-5 produced a higher number of eosinophil colonies in cultures compared to cells taken from wt mice 21 days after allergen exposure (Fig. 5).

Figure 5.

 Total number of eosinophil colonies after culture of CD34+ bone marrow cells from IFN-γR KO mice and wild type controls 21 days after OVA exposure. IFN-γR deficient mice display more eosinophil colonies than wild type mice (P < 0.05). Data are shown as mean ± SEM from three independent experiments where cells from 9 to 10 animals were pooled from each group.


The main finding in this study is that prolonged airway eosinophilia in a Th2 driven environment (IFN-γR KO Mice) is at least partly depended upon continuous de novo eosinophil production. Importantly, this is attenuated by the in vivo blockade of IL-5, both in the airways, as well as in the major eosinophilopoietic organ, the BM. The enhanced eosinophilopoiesis in the BM was confirmed by enhanced colony formation in CD34-cells harvested from the BM. Overall, the above findings argue that a Th2 establish environment is associated with prolonged eosinophilopoiesis.

It is known that IFN-γR KO mice have a very prolonged airway eosinophilia after a brief period of exposure to allergen, measurable up to 30 days [26]. However, these data originated from lung tissue quantifications alone. Our current data extend this previous finding to prove that the number of eosinophils also increases in the BM, under these circumstances, implying that an increased eosinophil production may be present. This probably led to the increased blood eosinophilia we also observed in IFN-γR KO after 3 weeks. However here, there is a discrepancy with the original study by Coyle et al., [26] as they did not observed differences in blood eosinophilia. This difference can be attributing to the different exposure protocol used in the two studies, and the methods of evaluating blood eosinophils. Coyle et al. expose their animals once with aerosol OVA while we exposed them three times with intranasally OVA inducing probably a stronger eosinophilic inflammation.

To document, the de novo production of eosinophils, BrdU was injected late after allergen, as a marker of mitosis. Importantly, more than 50% of BAL granulocytes were BrdU positive 3 weeks after the last allergen exposure. As the BrdU was injected 2 weeks after the last allergen exposure, these findings confirm that the precursors had undergone at least one mitosis during the last week prior to the cell harvest leading to BrdU-labelled eosinophils. Importantly, 96% of the granulocytes in the BAL were eosinophils as detected by Giemsa staining, telling us that almost all new granulocytes are eosinophils.

A main characteristic of IFN-γR KO mice has previously been shown to be a sustained capacity of the lung cells to a produce Th2 cytokines, including IL-5. We have confirmed this finding by detecting increased amount of IL-5 in serum 21 days after the last allergen exposure. IL-5 is crucial for eosinophils as mice lacking IL-5 fail to develop airway eosinophilia after allergen sensitization and exposure [27]. Furthermore, wt mice show a reduction in eosinophilic response after treatment with a neutralizing anti-IL-5 Ab [28]. IL-5 is both a survival factor for eosinophils, acting mainly by reducing their apoptosis [29], and therefore prolonged eosinophil survival has been argued to be a main mechanism for prolonged eosinophilic inflammation. However, IL-5 is also a key factor stimulating eosinophilopoiesis, and therefore prolonged IL-5 production may influence both of these aspects of the life of the eosinophil. As IL-5 controls the terminal differentiation of CD34+ progenitors to eosinophils [28], and is involved in active release of eosinophils to the circulation from the BM compartment [30, 31], we hypothesized that the BM may be involved in the prolonged eosinophilia observed in IFN-γR KO mice.

The extended activation of the BM after allergen exposure in the IFN-γR KO mice was also documented by enhanced IL-5 dependent eosinophil colony formation in CD34+ progenitors cells harvested from the BM 21 days after the last airway allergen exposure. In wt mice, it has been shown that allergen causes a transient increase in BM eosinophil colony formation, which disappears already after 3 days [32]. The relevance of this model to human disease can be argued, as human studies have shown that eosinophil colony-forming activity is increased in both BM and PB after allergen exposure and during asthma exacerbations [33, 34]. Persistent allergic inflammation is associated with increased number of circulating CD34+ cells in atopic patients, and in bronchial mucosa during asthma or in the nasal mucosa of patients with allergic rhinitis [35]. Therefore, the presence of new eosinophils in the airways may be a result of traffic of new cells from the BM, but also to some degree from local proliferation of progenitors in the airway tissue [21]. In an earlier study in our laboratory investigating the role of IL-5 specifically in eosinophilpopoiesis and CD34 progenitor’s cells, we have shown that a large fraction of eosinophils in BAL still express CD34 in their surface arguing for increased released from BM. Blockage of IL-5 resulted in decrease of eosinophilopoiesis in BM by inhibit CD34 cells to mature further into eosinophils. Even further we have recently showed that both CD4 and CD8 cells, a main source of IL-5, regulate BM eosinophilopoiesis [36, 37]. All above argue very strongly and confirm the importance of IL-5 in eosinophilopoiesis.

To elucidate in more detail the role of IL-5 in the current model of prolonged eosinophilia, we gave a group of animals a neutralizing anti-IL-5-antibody, together with BrdU during the last 6 days prior to cell harvest (2 weeks after the last allergen exposure). We found that newly produced eosinophils in BAL were substantially reduced. Two previous studies have argued that the principal site of action of IL-5 being the BM, where it stimulates the production of eosinophils [27, 28]. Consequently, anti-IL-5 given systemically (i.p.) is more potent than an anti-IL-5 antibody given locally in the airways in reducing airway eosinophilia, probably due to a key inhibitory effect being exerted in the BM. Furthermore, systemic but not local airway reconstitution of IL-5-production in IL-5KO mice reconstitutes allergen-induced airway eosinophilia in parallel with an increasing number of eosinophils in the BM. Therefore, a major site of effect of anti-IL-5 treatment is most likely localized to the BM, supporting the concept that the prolonged eosinophilia in IFN-γR KO mice is at least partly dependent on eosinophilopoiesis in the BM in these animals. However, we acknowledge that the anti-IL-5 treatment in the present study may also have reduced the allergen-induced airway eosinophilia by shortening eosinophil survival in the airways [38, 39] as approximately 50% of the granulocytes in the BAL which were BrdU negative also significantly decreased after anti-IL-5 treatment.

An interest attempt could be to speculate further into the underlying mechanisms. The Th1/Th2 hypothesis is now studied deeper, moved from the cytokine expression level into the level of transcription factors. Th1 cells characterized by the expression of T-bet while Th2 cells characterized by the expression of GATA-3 [40]. It has been shown now that during Th1 and Th2 differentiation pathways, Th2 differentiation inhibits Th1 commitment and vice versa. The picture is becoming even more complicated with the recent discovery of new T-cell phenotypes T-regulatory cells; expressing FOXP3 and Th17; expressing the transcription factor (ROR)gammat [14, 41]. We can hypotheses that in wild mice one of the above mechanisms is involved in the regulation of eosinophilic inflammation, a regulation that is incomplete in IFN-γ KO mice. In IFN-γ KO, we also found increased number of lymphocytes at day 21, which argues that the hypothesis is relevant. Especially, we recently have shown that both CD4 and CD8 can regulate BM eosinophilopoiesis [37].

The importance of the Th2 environment in allergic eosinophilic inflammation was also shown both in mice and in human by IFN-γ administration. In mice treated with IFN-γ prior or during induction of allergic inflammation, a protective effect has been documented [42], while recombinant IFN-γ has some, but weak, beneficial effects on the degree of circulating eosinophils in patients with atopic dermatitis as well as in asthmatic patients [43, 44]. In an earlier study, we have shown that PB mononuclear cells from atopic dermatitis children produce more IL-4 and less IFN-γ, a difference that was partially reversed if cells were pretreated with exogenous IFN-γ [45]. A recent study has also revealed a protective role for IFN-γ in anaphylaxis. Nieuwenhuizen et al. [46] have shown recently in a mouse model of anaphylaxis that IFN-γ play a direct role in preventing fatal anaphylactic reactions by inhibiting mast cells degranulation. More specifically, our data complement previous studies concerning the role of the BM in a Th2 environment. Rais et al. [47] have shown that IL-12 inhibits eosinophil differentiation of BM progenitors cultured with IL-5 by 40–45%, an effect that was mediated by IFN-γ, partly explaining our results. However, recent studies argue that the picture is not yet completed and the role of IFN-γ especially in severe asthma is not yet perfectly clear. Recently, Truyen et al. [48] studied Th1/Th2 cytokine mRNA profile in sputum of asthmatics and healthy subjects. They found that sputum mRNA levels of IL-4, IL-5 and IL-13 were higher in asthmatic and correlated with eosinophilia, but they did not differ between mild and severe patients. Opposite IFN-γ mRNA expression was higher in controls compared to asthmatic but more important it was increased in patients with moderate to severe asthma compared to mild one, proposing that IFN-γ can indicate asthma severity [48].

We used a short-term exposure model which showed persistent eosinophilia after 3 weeks, attempting to mimic an ‘exposure exacerbation’ in a Th2 established environment. There is a discrepancy between results of experimental animal’s models and humans. In a mouse model, we have an extended eosinophilic inflammation and not a characteristic asthma phenotype. However, it is a valuable model to investigate the underlying mechanisms of eosinophilic inflammation. Even more as we described earlier, eosinophils represent a main inflammatory cell in asthma [48].

In conclusion, our study has shown that in a model of a Th2-driven prolonged airway eosinophilia, there is a long-term increased number of newly produced eosinophils in BAL, blood and BM, together with enhanced responsiveness of BM CD34+ progenitor cells to IL-5. These data together strongly argue that eosinophilopoiesis is involved in prolonged airway eosinophilic inflammation, by continuous supply of new eosinophils. Consequently, we suggest that eosinophilopoiesis is not only involved in the initiation of allergic eosinophilic inflammation, but also in the maintenance of airway eosinophilia, which is regulated by the lack of IFN-γ, and dependent on IL-5, as anti IL-5 rapidly reduces both BM and the number of BAL eosinophils.


The authors are grateful to Prof. N.Y. Lycke (Department of Clinical Immunology, University of Göteborg, Sweden) for kindly providing us with the IFN-γ receptor knockout mice, to Prof. Bengt-Eric Skoogh for valuable comments on the manuscript, and to Mrs Carina Malmhäll for technical assistance. This work was supported by the Swedish Heart and Lung Foundation and the Verbal Foundation. Apostolos Bossios received grant and financial support from European Academy of Allergy and Clinical Immunology (EAACI) and from ‘EMPIRIKION foundation’, Athens, Greece.