Chemotaxis of bone marrow-derived eosinophils in vivo: A novel method to explore receptor-dependent trafficking in the mouse

Summary

Here we describe a novel method via which ex vivo cultured mouse bone marrow-derived eosinophils (bmEos) can be adoptively transferred into recipient mice in order to study receptor-dependent recruitment to lung tissue in vivo. Intratracheal instillation of recombinant hCCL24 prior to introduction of bmEos via tail vein injection resulted in a ~four-fold increase in Siglec F^pos/CD11c^neg eosinophils in the lungs of eosinophil-deficient ΔdblGATA recipient mice compared with controls. As anticipated, bmEos generated from CCR3-gene-deleted mice did not migrate to the lung in response to hCCL24 in this model, indicating specific receptor dependence. BmEos generated from GFP^pos Balb/c mice responded similarly hCCL24 in vitro and were detected in lung tissue of BALB/c wild-type as well as BALB/c ΔdblGATA eosinophil-deficient recipient mice, at ~four-fold (at 5 h post-instillation) and ~three-fold (at 24 h post-instillation) over baseline respectively. Comparable results were obtained with GFP^pos C57BL/6 bmEos responding to intratracheal hCCL24 in C57BL/6 ΔdblGATA recipient mice. The use of ex vivo cultured bmEos via one or more of these methods offers the possibility of manipulating bmEos prior to transfer into a wild-type or gene-deleted recipient host. Thus, this chemotaxis model represents a novel and robust tool for pharmacological studies in vivo.

Introduction

The recruitment of eosinophils from the circulation into the lungs and airways is a hallmark of allergic lung inflammation. During allergic airway inflammation, eosinophils are recruited to the lungs by gradients of lipid mediators, such as PGD₂ [1] and chemokines such as eotaxins, which are produced by airway epithelial cells under the influence of other pro-inflammatory cytokines [2, 3].

Recently Dyer et al. demonstrated that ex vivo cultured bone marrow-derived eosinophils (bmEos) migrate specifically in response to mouse eotaxin-1 in vitro [4]; the importance of this signaling pathway in human eosinophilic disease and in mouse models has already been established [5]. The eotaxin sub-family comprises eotaxin-1/hCCL11, eotaxin-2/hCCL24 and eotaxin-3/hCCL26 in humans [6] and eotaxin-1/mCCL11 and eotaxin-2/mCCL24 in mice [7]. Structural similarity within the human eotaxin family is relatively low: sequence homologies of only 39% between hCCL24 and hCCL11 [8], 37% between hCCL26 and hCCL11, and 34% between hCCL26 and hCCL24 have been described [9]. Interestingly, mCCL24 is most homologous to hCCL24 (59.1%), while only 38.9 and 38.2% homologous to mCCL11 and hCCL26 respectively [10]. Nevertheless, a common feature of all eotaxins is that they act exclusively via the CC-chemokine receptor CCR3.

Human CCL24 is further characterized as a highly selective CCR3-receptor agonist that induces shape change, chemotaxis, and the release and synthesis of pro-inflammatory mediators in both human and mouse eosinophils [11, 12]. In both humans and mice, CCL24 was shown to cooperate with other cytokines, including other eotaxins [2, 13], interleukin (IL)-5 [11, 14] and IL-13 [7, 15, 16] to promote accumulation of eosinophils in the tissue. In humans, hCCL24 is expressed in the lungs of both atopic and non-atopic asthmatics [17] and a delayed increase in mCCL24 levels was detected in a mouse model 24 h after allergen-challenge [18].

In the past, eosinophils were mainly thought to be end stage effector cells in inflammatory and allergic processes, controlled by T-cell responses, but recent findings have defined a more complex role for these cells [19]. Studies in mouse models suggest that eosinophils may also play an important role early in the development of allergic asthma [20]. Moreover, eosinophils are now considered multifunctional cells involved in tissue homeostasis [21] and modulation of adaptive immune responses and innate immunity to certain microbes [22]. Thus, eosinophils are among the most enigmatic cells of our immune system and their true immunological roles still need to be elucidated.

Previous in vivo migration studies utilize the more complex allergen-exposed models [23] or the in vivo migration of endogenous eosinophils into the BAL fluid in response to an exogenous chemotactic stimulus [12]. These methods have only limited flexibility by nature. Thus, it is of primary importance to establish new models for studying eosinophil migration in vivo. The principal intent of this work was to create a new model via which ex vivo cultured mouse bone marrow eosinophils (bmEos), can respond appropriately via the crucial CCR3 signaling pathway when adoptively transferred into mouse peripheral circulation in vivo.

Results

Trafficking and clearance of mature bmEos in eosinophil-deficient recipient mice

In order to optimize the recovery rates of the present adoptive transfer protocol, mature Balb/c bmEos (10⁷ cells [24, 25]/200 µl PBS) were injected intravenously into eosinophil-deficient ΔdblGATA mice and their homeostatic recruitment into the spleen was assessed by flow cytometric analysis. BmEos located in the tissue were identified as Siglec F^pos granulocytes in a higher side scatter region [26]. As shown in Fig. 1, the recovery rates were similar to those described in the literature [27]; the percentage of Siglec F^pos granulocytes recovered remained stable up to 24 h after the transfer, but was significantly decreased 48 h post-injection. Recovery rates of 13.4 ± 0.6% Siglec F^pos granulocytes were observed 5 h to 24 h after the transfer, while rates of 6.2 ± 1.7% were found after 48h. Thus, we chose to focus on a 5 h to 24 h post-injection time interval as appropriate to explore eosinophil migration.

Figure 1. Frequency of bmEos in spleens of eosinophil-deficient ΔdblGATA recipient mice.

Mature bone marrow-derived eosinophils (bmEos) were transferred into ΔdblGATA mice via tail vein injection. Spleens from recipient mice were collected at different time-points post-injection and recovery was determined by flow cytometry. Granulocytes in spleen single-cell suspensions were identified by forward and side scatter properties. The frequency of bmEos is expressed as percentage of Siglec F^pos cells within the granulocyte gates. Each symbol represents an individual mouse and data are presented as mean ± SEM of 3 – 9 mice/time point. Data shown are pooled from 6 independent experiments. *p < 0.05, **p < 0.01, Student’s t test.

In vivo chemotaxis protocol

Our essential protocol is outlined in Fig. 2. Bone marrow cells were collected from Balb/c donor mice and differentiated into mature eosinophils using an ex vivo cell culture system [4]. Bone marrow-derived eosinophils (bmEos) were collected on day 10 and 10⁷ cells were transferred into eosinophil-deficient ΔdblGATA mice via tail vein injection. Immediately prior to the cell transfer, recipient mice were either left untreated or were anesthetized and received intratracheal instillation of hCCL24 (5 µg/50 µl sterile saline) [12]. Human CCL24 is a chemotactic factor for eosinophils with a similar chemotactic potency to hCCL11 [8, 28]. Borchers et al. [12, 29] confirmed the potent chemoattractant activity of hCCL24 on mouse eosinophils in vitro [12, 29], while Ochkur et al. [12] found that hCCL24 was a superior chemoattractant for mouse eosinophils compared to mCCL24 in studies performed in vivo. Lungs from recipient mice were collected 5h or 24h after the transfer and lung single-cell suspensions were prepared. Specific recruitment in vivo was determined by flow cytometric analysis of lung single-cell suspensions. Recruited bmEos were identified as Siglec F^pos/CD11c^neg cells in whole lung tissue [26]. Additionally, cytospins were prepared from lung single-cell suspensions and stained with Diff-Quik.

Figure 2. Schematic representation of the in vivo chemotaxis protocol.

(A) Bone marrow cells from Balb/c donor mice are differentiated into mature eosinophils as described by Dyer et al. [4]. (B) Directed migration into the lung tissue is induced by intratracheal instillation of hCCL24 and (C) bone marrow-derived eosinophils (bmEos) are transferred into eosinophil-deficient ΔdblGATA mice via tail vein injection. (D) Lungs from the recipient mice are collected at different time-points post-instillation and % eosinophils are determined by flow cytometric analysis of lung single-cell suspensions. Migrated bmEos are identified as live Siglec F^pos/CD11c^neg cells.

Human CCL24 elicits chemotactic responses of Balb/c bmEos both in vitro and in vivo

In vitro chemotactic responses of Balb/c bmEos (day 10–11) were evaluated as shown in Fig. 3A. BmEos were placed in the upper wells of a 96-well chemotaxis chamber and were allowed to migrate towards serial dilutions of hCCL24 (1 nM – 1 µM) for 3 hours. BmEos migrated towards hCCL24 in a concentration-dependent fashion with a maximum chemotactic index (CI) = 4, which was observed at the highest concentration of hCCL24 (1 µM).

Figure 3. Frequency of migrated Balb/c bmEos in the lung tissue of eosinophil-deficient recipient mice as determined by flow cytometry.

(A) BmEos (10⁵/100 µl) were placed into the upper wells of a 96-well chemotaxis plate and were allowed to migrate towards hCCL24 or vehicle controls for 3 h. Migrated cells were enumerated by flow cytometry. In vitro migration is expressed as the chemotactic index (CI); the data are shown as mean ± SEM from 9 separate cultures. (B–E) hCCL24 (5 µg/50 µl) was instilled via the trachea into ΔdblGATA mice and Balb/c bmEos were transferred via tail vein injection (10⁷cells/mouse). Lungs of recipient mice were harvested (B, D) 5 h or (C, D) 24 h after the transfer for flow cytometric analysis. (B, C) The frequency of migrated bmEos is expressed as % Siglec F^pos/CD11c^neg cells among vital Live-Dead negative cells and are shown as mean ± SEM of 11 – 18 mice/condition. Data shown are pooled from 4–6 experiments. ***p < 0.005, Student’s t test. (D, E) Chemotactic in vivo responses were also expressed as the chemotactic index and shown as mean ± SEM of 11 – 18 samples pooled from 4 (D) and 6 (E) independent experiments; *p < 0.05, **p < 0.01, Student’s t test.

The ability of hCCL24 to elicit migration of adoptively transferred Balb/c bmEos into lung tissue in vivo was examined in ΔdblGATA mice. Exogenous hCCL24 (5 µg in 50 µl of sterile saline) was instilled into the trachea of recipient mice and Balb/c bmEos (10⁷cells/mouse) were transferred via tail vein injection immediately thereafter. Lungs of recipient mice were harvested 5 h or 24 h post-injection and lung single-cell suspensions were stained for flow cytometric analysis and visual inspection. For FACS analysis migrated bmEos were identified as live Siglec F^pos/CD11c^neg cells. The fraction of bmEos recovered from lung tissue of untreated mice was 1.28 ± 0.29% at 5h after the eosinophil transfer (Fig. 3B) which decreased to 0.26 ± 0.03% at 24 h after transfer (Fig. 3C), most likely due to rapid clearance. Intratracheal administration of hCCL24 resulted in a significant increase in bmEos in the lung tissue of treated mice compared to untreated control mice. This translates to a chemotactic index (CI) = 3.6 detected at 5 h after eosinophil transfer (Fig. 3D) and a CI = 3 at 24 h after eosinophil transfer (Fig. 3E). Similar results were obtained by visual inspection of Diff-Quik stained cytospins (Fig. 4A – D). The percentage of eosinophils detected in lungs of untreated mice was 1.4 ± 0.19% at 5h (Fig. 4A) and 0.39 ± 0.07% at 24 h (Fig. 4B) after the transfer. Intratracheal hCCL24 elicited a chemotactic in vivo response with an average CI = 4.6 at 5 h (Fig. 4C) and an average CI = 3.3 at 24 h (Fig. 4D). In this model, instillation of hCCL24 into the trachea induced bmEos recruitment into the lung tissue of recipient mice but not into their airways. Interestingly, at both time-points (5 h and 24 h), eosinophil cell counts in BAL fluid from hCCL24-treated mice remained the same as those from untreated control mice (data not shown). There are many potential explanations for this observation. The most likely is that that it takes the bmEos longer to migrate into the airways, and to be detected in the BALF. Alternatively, priming with IL-5 may be required to promote transmigration [30].

Figure 4. Frequency of migrated Balb/c bmEos in lung tissues of eosinophil-deficient recipient mice as determined by visual inspection.

Recombinant hCCL24 was instilled via the trachea of ΔdblGATA mice and Balb/c bmEos were transferred via tail vein injection (10⁷cells/mouse). Lungs of recipient mice were harvested (A, C) 5 h or (B, D) 24 h post-injection. The percentage of bmEos was determined by visual inspection of Diff-Quik stained cytospins of lung single-cell suspensions (500 cells per slide). Each symbol represents an individual mouse and data are presented as mean ± SEM of 9 – 19 mice/condition. Data shown are pooled from 4–6 experiments. ***p < 0.005, Student’s t test. (C, D) Chemotactic in vivo responses were expressed as the chemotactic index (CI). The data are shown as mean ± SEM of 9 – 19 samples pooled from 4 (C) and 6 (D) independent experiments; **p < 0.01, Student’s t test.

Correlation between FACS analysis and visual inspection

To verify the relationship between FACS analysis and visual inspection of Diff-Quik stained slides, correlation analyses of in vivo chemotactic indices (CI) were conducted. CI values obtained by FACS analysis (CI_FACS) were not significantly different from data obtained by visual inspection (CI_Diff-Quik) (p=0.276; Fig. 5A). Linear regression analysis yield a significant positive correlation between data from both methods with a correlation coefficient R = 0.79 (R² = 0.62; Fig. 5B). The Bland-Altman plot, in which the average CI [(CI_Diff-Quik + CI_FACS)/2] was plotted against the difference between CI_Diff-Quik and CI_FACS yield a positive mean bias of 0.615 (SD = 0.856; 95% CI = −0.0097 to 1.2392), and limits of agreement at −1.064 (95% CI = −2.1451 to 0.0179) and 2.292 (95% CI = 1.2115 to 3.3746). The scatter on the Bland-Altman plots was distributed randomly, without signs of systematic error (Fig. 5C). This shows a good agreement between results obtained from FACS analysis and visual inspection, without any data points outside 1.96 SD from the mean.

Figure 5. Correlation between flow cytometric analysis and visual inspection.

To confirm the agreement between flow cytometric analysis and the visual inspection method, correlation analyses of in vivo chemotactic indices (CI) were conducted. (A) There was no significant difference between in vivo chemotactic indices obtained by FACS (CI_FACS) and visual inspection (CI_Diff-Quik). Data are presented as mean ± SEM from 10 independent experiments; p=0.275, CI values of hCCL24-treated recipients versus un-treated controls for Student’s t test. (B, C) Correlation of CI_FACS and CI_Diff-Quick was displayed as (B) a scatter plot and (C) as a Bland-Altman plot (mean bias = 0.615, limits of agreement = −1.0636, to 2.2931).

Confirmation of the CCR3-specific chemotaxis of Balb/c bmEos

To confirm that our observations are due to selective CCR3-receptor activation, bmEos were generated from CCR3-deficient donor mice and transferred into eosinophil-deficient recipients. CCR3-deficient progenitors responded to cytokine-stimulation and differentiated and expanded in a similar manner as wild type progenitors, both generating bmEos (Fig. 6A). As expected, CCR3−/− bmEos did not migrate towards hCCL24 in a standard in vitro chemotaxis assay (Fig. 6B).

Figure 6. Human CCL24 induced eosinophil in vivo chemotaxis in a CCR3 receptor-dependent manner.

(A) CCR3-deficient bmEos differentiated similar to WT bmEos, but (B) did not respond towards hCCL24 in vitro. In vitro migration was expressed as the chemotactic index (CI); the data are presented as mean ± SEM from 6 separate cultures. (C–F) hCCL24 (5 µg/50 µl) was instilled via the trachea of ΔdblGATA mice and CCR3-deficient or WT bmEos were transferred via tail vein injection (10⁷cells/mouse) into the indicated recipients. Lungs of recipient mice were harvested at 24 h after the transfer for standard analysis. The frequency of migrated bmEos is expressed as (C) the percentage of live Siglec F^pos/CD11c^neg cells or (E) as the percentage of bmEos in the lung single-cell suspension. Data are presented as mean ± SEM of 6–7 mice/condition and are pooled from 2 independent experiments. **p < 0.01, ***p < 0.005, Student’s t test. (D, F) Chemotactic in vivo responses were expressed as the chemotactic index (CI) and presented as mean ± SEM of 6 – 7 samples pooled from 2 independent experiments.

For in vivo experiments, hCCL24 was instilled via the trachea to ΔdblGATA mice and either CCR3-deficient or wild type bmEos were transferred via tail vein injection (10⁷ cells/mouse). Lungs of recipient mice were harvested at the 24 h time point for standard analysis. The frequency of live Siglec F^pos/CD11c^neg cells obtained from lungs of control (no hCCL24) recipient mice was 0.31 ± 0.06% for CCR3-deficient bmEos and similarly, 0.31 ± 0.03% for wild type bmEos (Fig. 6C). Instillation of hCCL24 induced a chemotactic in vivo response in wild type bmEos (CI = 2), but had no impact on the recovery of CCR3-deficient bmEos (Fig. 6D). Similar results were obtained with visual inspection of Diff-Quik stained samples (Fig. 6E and F).

The use GFP^pos bmEos facilitates flow cytometric identification of migrated cells

To further facilitate the flow cytometric identification of migrated eosinophils, bmEos differentiated from bone marrow cells of GFP-positive CByJ.B6-Tg(UBC-GFP)30Scha/J mice were used [31].

On day 10 of culture, bmEos generated from CByJ.B6-Tg(UBC-GFP)30Scha/J mice were 95% GFP^pos (Fig. 7A) and fully differentiated into functionally competent eosinophils, which responded similarly to wild type bmEos in vitro (Fig. 7B). To assess in vivo migration of GFP^pos bmEos, exogenous hCCL24 was instilled via the trachea of ΔdblGATA mice prior to the adoptive transfer, and lungs of recipient mice were harvested 24 h post-injection for standard analysis. Recovered GFP^pos bmEos were identified as live Siglec F^pos/GFP^pos cells and the frequency of GFP^pos bmEos was expressed as percentage of total live cells (Fig. 7C). The fraction of GFP^pos bmEos recovered from untreated mice was slightly lower (0.13 ± 0.03%) but comparable to that achieved for wild type bmEos (Fig. 7D) and administration of exogenous hCCL24 likewise resulted in a comparable chemotactic in vivo response with a chemotaxis index (CI) = 2.8 (Fig. 7E and F). Similar results were obtained with visual inspection of Diff-Quik preparations (500 cells per slide; Fig. 6G and H). Thus, with respect to the present data, the responses of GFP^pos bmEos generated from CByJ.B6-Tg(UBC-GFP)30Scha/J mice are indistinguishable from those of wild type bmEos in vitro and in vivo, and represent another useful tool for the study of eosinophil in vivo trafficking.

Figure 7. The adoptive transfer of GFP^pos bmEos facilitates flow cytometric identification of migrated cells.

(A) BmEos generated from CByJ.B6-Tg(UBC-GFP)30Scha/J mice were 95% GFP^pos and differentiated similar to WT bmEos. A representative histogram of flow cytometric analysis of GFP expression in bmEos from day 10 cultures is shown in the insert. (B) GFP^pos bmEos exhibited in vitro chemotaxis towards hCCL24 similar to WT bmEos. In vitro migration was expressed as the chemotactic index (CI); the data are presented as mean ± SEM from 4 separate cultures. (C–H) hCCL24 was instilled via the trachea of ΔdblGATA recipient mice and GFP^pos bmEos were transferred via tail vein injection (10⁷cells/mouse). Lungs of recipient mice were harvested 24 h post-injection for standard analysis. All flow cytometry analyses were performed on (C) vital Live-Dead negative cells and (D) migrated GFP^pos bmEos were identified as GFP^pos/Siglec F^pos cells. (E, G) The frequency of migrated bmEos is expressed as (E) the percentage of live GFP^pos/Siglec F^pos cells or (G) as the percentage of bmEos in the lung single-cell suspension. Data are presented as mean ± SEM of 8–12 mice/condition and are pooled from 3 independent experiments. *p < 0.05, ***p < 0.005, Student’s t test. (F, H) Chemotactic in vivo responses were also expressed as the chemotactic index (CI) and shown as mean ± SEM of 8 – 12 samples pooled from 3 independent experiments; *p < 0.05; **p < 0.01, (Student’s t-test.

Transfer of GFP^pos bmEos allows study of eosinophil tracking in wild type recipient mice

In order to explore eosinophil chemotaxis under more physiological conditions, hCCL24 was instilled intratracheally into Balb/c (eosinophil-sufficient) wild type recipient mice and GFP^pos bmEos were transferred via tail vein injection (10⁷cells/mouse). Lungs of recipient mice were harvested at time point 5 h or 24 h post-injection and single-cell suspensions were stained for FACS analysis. The number of migrated Siglec F^pos/GFP^pos bmEos was expressed as percentage of total live cells. In mice that received the bmEos only, recovery was 0.94 ± 0.23% at the 5 h time point (Fig. 8A) and decreased to 0.32 ± 0.9.7% at 24h after the transfer (Fig. 8B). Interestingly, at 5 h after the transfer, a significant increase in the frequency of hCCL24- recruited GFP^pos bmEos was detected (CI = 3.8; Fig. 8C), whereas at 24 h after the transfer, no significant increase in GFP^pos bmEos recovery was observed (Fig. 8D).

Figure 8. The adoptive transfer of GF^pos bmEos allows eosinophil tracking in Balb/c WT mice.

(A–D) Recombinant hCCL24 was instilled via the trachea of Balb/c WT recipient mice and GFP^pos bmEos were transferred via tail vein injection (10⁷cells/mouse). The lungs of recipient mice were harvested (A, C) 5 h or (B, D) 24 h post-injection for FACS analysis. Migrated GFP^pos bmEos were expressed as (A, B) the percentage of live Siglec F^pos/GFP^pos cells in the lung single-cell suspension and the data are presented as mean ± SEM of 8–9 mice/condition and are pooled from 3–4 independent experiments. **p < 0.01, Student’s t test. (C, D) Chemotactic in vivo responses were expressed as the chemotactic index (CI) and shown as mean ± SEM from 8 – 9 samples pooled from 3–4 independent experiments; *p < 0.05, Student’s t-test.

The in vivo chemotaxis model on the C57Bl/6 background

To explore the impact of different genetic backgrounds on the chemotactic responsiveness of bmEos in vivo and in vitro, bone marrow cells from GFP+ C57BL/6-Tg(UBC-GFP)30Scha/J donor mice were collected and differentiated into bmEos as described previously [32]. On day 13 to 14 of culture, the fully differentiated C57Bl/6 GFP^pos bmEos (Fig. 9A) showed higher in vitro responses towards hCCL24 than did similarly matured Balb/c GFP^pos bmEos. The chemotactic index (CI) was 8 for C57Bl/6 GFP^pos bmEos from day 13–14 cultures (Fig. 9B), as compared to only 3–4 for Balb/c GFP^pos bmEos from day 10–11 cultures (Fig. 7B). To assess the in vivo responsiveness, mature C57Bl/6 GFP^pos bmEos were adoptively transferred into C57Bl/6 ΔdblGATA mice and directed migration into the lung tissue was induced with hCCL24 as described above. Lungs from recipient mice were collected 5h after the transfer and prepared for standard analysis. For FACS analysis, migrated bmEos were identified as live Siglec F^pos/GFP^pos cells. The frequency of GFP^pos bmEos in the lung tissue at baseline was 1.16 ± 0.24% (Fig. 9C) in mice that only received the bmEos. Administration of exogenous hCCL24 elicited a chemotactic in vivo response with a chemotactic index (CI) = 3.1(Fig. 9D), and similar results were obtained with visual inspection of Diff-Quik preparations. The percentage of eosinophils was 1.7 ± 0.24% in untreated mice and 4.68 ± 0.99% in mice that received intratracheal hCCL24 (Fig. 9E), representing a CI = 3 (Fig. 9F). As such, we conclude that the in vivo chemotaxis model presented here yields consistent results in both Balb/c and C57Bl/6 mouse strains.

Figure 9. Frequency of GFP^pos C57Bl/6 bmEos in lung tissue of eosinophil-deficient recipients.

(A) BmEos generated from C57BL/6-Tg(UBC-GFP)30Scha/J mice fully differentiated into mature eosinophils similar to WT bmEos. (B) Recombinant hCCL24 elicited a strong chemotactic response in GFP^pos C57Bl/6 bmEos in vitro. In vitro migration was expressed as the chemotactic index (CI) and presented as mean ± SEM from 4 separate cultures. (C–F) For in vivo experiments, hCCL24 was instilled via the trachea of C57BL/6 ΔdblGATA mice and GFP^pos C57Bl/6 bmEos were transferred via tail vein injection (10⁷cells/mouse). Lungs of recipient mice were harvested 5 h after the transfer for standard analysis. (C) All flow cytometric analyses were performed on vital Live-Dead negative cells and migrated GFP^pos bmEos were identified as GFP^pos/Siglec F^pos cells. (C, E) The frequency of migrated bmEos is expressed as (C) the percentage of live GFP^pos/Siglec F^pos cells or (E) as the percentage of bmEos in the lung single-cell suspension. Data are presented as mean ± SEM of 6 mice/condition and are pooled from 3 independent experiments. *p < 0.05, Student’s t test. (D, F) Chemotactic in vivo responses were also expressed as the chemotactic index and presented are mean ± SEM of 6 samples pooled from 3 independent experiments;*p < 0.05, Student’s t test.

Discussion

Here we describe a novel mouse model that provides the unique opportunity to explore eosinophil directed migration in vivo. This technique is based on the intravenous adoptive transfer of ex vivo differentiated and expanded bone marrow-derived eosinophils (bmEos) [4] into wild type or eosinophil-deficient ΔdblGATA recipient mice. As a chemotactic stimulus, recombinant hCCL24 [12, 28] was administered via the trachea of recipient mice and recruitment was evaluated by FACS analysis or visual inspection of lung single-cell suspensions. Although this method does not permit us to evaluate specific migration efficiency, we observed that bmEos migrate to the lungs in response to hCCL24 in a receptor dependent fashion.

The highest recovery rates of bmEos in spleens of baseline recipient mice were observed at 5 h after the transfer; 24 h the percentage of bmEos had already declined, whereas at 48 h most of the bmEos have already been cleared. These data are in accordance with a recent study on eosinophil migration and survival in vivo, where mature eosinophils were adoptively transferred into naïve and Nocardia brasiliensis infected mice and the percentage of transferred eosinophils was estimated in several tissues [27].

Several in vivo models of eosinophil recruitment related to respiratory diseases have been presented by other laboratories. Most of these models are based on the more complex allergen-challenge [23, 33] or describe the recruitment of endogenous eosinophils into the BAL fluid in response to administration of exogenous chemokine [12, 30, 34, 35]. Several groups have shown that eotaxins are potent inducers of mouse and guinea pig eosinophil migration in vitro and in vivo [12, 29, 30, 36]. These methods are rapid and robust, but with a major limitation: endogenous eosinophils cannot be manipulated in order to examine specific signaling pathways. One of the main advantages of the technique presented here is that the bmEos are differentiated in an ex vivo culture system. This step offers the unique possibility of manipulating eosinophils with pharmacological tools, such as receptor-specific agonists, antagonists or kinase-specific inhibitors prior to the adoptive transfer or of producing bmEos from genetically modified mice. Of particular interest in this case, several characterized CCR3 antagonists cannot interact effectively with murine receptors [37]. As such, generation of bmEos in which the human receptor has been "knocked-in" to replace the murine counterpart would represent a further improvement in vivo migration model. Moreover, the introduction of synthetic small interfering RNA (siRNA) into the cultured cells could trigger highly efficient gene-silencing and thus, may provide the opportunity for probing the function of chemokine-receptor specific genes.

Further, this method also allows us to study the in vivo migration of adoptively transferred bmEos under physiologic, eosinophil-sufficient conditions as well as in an eosinophil-deficient environment in ΔdblGATA mice. Cells isolated from GFP^pos mice are commonly used in in vivo leukocyte trafficking studies [31]. Here we have shown that our ex vivo cell culture system generates large numbers of GFP^pos bmEos and that these cells migrate in a manner that is indistinguishable from GFP-negative wild-type bmEos.

In a previous publication, we found that eosinophil cultures with a purity approaching 100% can be generated from bone marrow from C57Bl/6 mice, but with the limitation that C57Bl/6 bmEos cultures do not expand as much as do the Balb/c bmEos [32]. Since the genetic background of experimental animals might affect inflammatory processes, we examined possible strain-specific differences in the chemotactic in vivo and in vitro response of Balb/c and C57Bl/6 bmEos. Interestingly, although C57Bl/6 bmEos respond more prominently to hCCL24 in the in vitro chemotaxis assay, we could not detect any significant phenotypic differences between C57BL/6 and BALB/c GFP^pos bmEos in response to hCCL24 when transferred into strain-matched ΔdblGATA mice. Therefore we conclude that the responses of C57BL/6 and BALB/c mice are indistinguishable in the present in vivo migration model in response to hCCL24, but further studies regarding the two strains’ in vivo behavior and responsiveness towards other chemoattractants need to be done.

In summary, we here describe a unique murine in vivo chemotaxis model that can be used to explore receptor-specific trafficking of bone marrow-derived eosinophils. The present model not only allows tracking of wild type eosinophil recruitment into the lung tissue, but also offers the possibility of generating large numbers of bmEos from receptor-specific knock-out mice and of transferring these bmEos into eosinophil-deficient ΔdblGATA mice. The differentiation of GFP^pos bmEos facilitates the flow cytometric detection of migrated cells in the lung tissue. Moreover, the use of GFP^pos bmEos enables the tracking of eosinophils under physiological conditions also in wild type recipients, since transferred bmEos can be readily distinguished from innate eosinophils. The present model represents a robust and flexible in vivo method to investigate eosinophil function on the Balb/c as well as C57Bl/6 background.

To conclude, the development of novel in vivo migration models is of primary interest to researchers studying eosinophil biology and effector function with respect to respiratory diseases. This unique technique allows the manipulation of assay parameters and therefore represents a very useful tool for pharmacological studies in vivo.

Material and methods

Mouse strains

Six- to 8-week-old BALB/c mice were purchased from the Division of Cancer Technologies (DCT, Rockville, MD). Four-week-old CCR3 receptor-deficient C.129S4-Ccr3^tm1Cge/J mice [3] and breeding pairs of the CByJ.B6-Tg(UBC-GFP)30Scha/J and C57BL/6-Tg(UBC-GFP)30Scha/J strains [31] were obtained from the Jackson Laboratory and maintained on-site. Eosinophil-deficient ΔdblGATA mice on the BALB/c and C57BL/6 backgrounds [38] were maintained on-site. This study was reviewed, and all protocols were carried out in accordance with the Institute's Animal Care and Use Committee Guidelines (Animal Study Protocol LAD 7E).

Generation of bone marrow-derived eosinophils (bmEos)

Bone marrow-derived eosinophils (bmEos) were differentiated ex vivo from unselected bone marrow progenitors using a well- defined cytokine regimen [4]. Our cultures yielded high-purity (90–100%) bmEos by day 10 (Balb/c) or day 13 (C57Bl/6) of culture as estimated by visual inspection of Diff-Quik stained cytospins.

Intratracheal instillation

Mice were sedated with intraperitoneal injection of ketamine (120 mg/kg; KetaVed, Vedco; Saint Joseph, MO) and xylazine (20 mg/kg; AnaSed^®, Lloyd Inc.; Shenandoah, IA) and manually restrained in an upright position on a rodent intubation stand (BioLite, BioTex Inc.; Houston, TX). Mice were inoculated with hCCL24 (5 µg/50 µl sterile saline, R&D Systems, Minneapolis, MN) using a standard ball-end gavage needle placed directly into the trachea.

Adoptive transfer of bone marrow-derived eosinophils

Fully differentiated bone marrow-derived eosinophils (bmEos) were adoptively transferred into recipient mice by tail vein injection. In brief, bmEos were collected on day 10 (Balb/c) or day 13 (C57Bl/6) of culture, washed, counted and suspended at 10⁷/200 µl in sterile PBS. Prior to the adoptive transfer, recipient mice were placed under a heat lamp for 2 min to promote vascular dilation. Recipient mice were restrained to allow access to the tail and 10⁷ bmEos were injected into the tail vein using a 28-gauge 1 ml - insulin syringe.

Collection of bronchoalveolar lavage (BAL) fluid

Bronchoalveolar lavage (BAL) fluid was collected from anesthetized recipient mice by instillation and withdrawal of 0.8 ml cold phosphate buffered saline with 0.1 % bovine serum albumin. Cytospin preparations were stained with Diff-Quik and the frequency of migrated bmEos was determined by visual inspection (500 cells/slide).

Spleen and lung cell suspensions

Single-cell suspensions of spleens from recipient mice were obtained as described previously [26]. Approximately 2–10 × 10⁷ cells were recovered from a ΔdblGATA mouse.

Lungs from recipient mice were perfused by injecting the right ventricle with 8 ml of perfusion media, PBS (KB Medical, Columbia, MD) with 10 mM EDTA (Quality Biologicals, Gaithersburg, MD). Lungs were minced and incubated in 8 ml digestion media: RPMI 640 supplemented with 20 µg/ml DNAse I (Roche Diagnostics, Indianapolis, IN), 5% FBS and 2 mg/ml collagenase D (Roche Diagnostics, Indianapolis, IN). The contents were stirred for 45 min at 37 °C, then 3 ml of fresh digestion media were added and the contents were incubated for another 45 min. After incubation, EDTA was added to a final concentration of 10 mM and samples were incubated for 5min at room temperature. The contents of the tubes were strained through a 70 µm cell strainer. The strainer was washed with 5 ml Wuerzburg buffer (PBS/0.3% BSA/5 mM EDTA/20 µg/ml DNAse I), cells were spun and pellets were resuspended in 1–3 ml of ACK lysing buffer (Lonza, Walkersville, MD) and incubated for 3–5 min at room temperature. Wuerzburg buffer was added and cells were washed with HBSS (Lonza, Walkersville, MD). Lung single-cell suspensions were counted using a hemacytometer. Approximately 1–3 × 10⁷ cells were recovered from lungs from wild-type naïve mice.

Slide preparation and differential count

Differential counting was accomplished by generating cytospin slides from bone marrow, bone marrow-derived eosinophils, spleens and lung cell suspensions as described previously [26]. Five hundred cells from each slide were counted under a 64× oil immersion objective and eosinophils are reported as percent of total cells. Eosinophils were detected based on the pink to red staining of their specific granules.

Flow cytometric analysis

Spleen or lung cells were resuspended in HBSS (Lonza, Walkersville, MD) at 10⁶/ml and stained with anti-Siglec F-PE (clone E50-2440, BD Pharmingen, San Jose, CA), anti-CD11c-Alexa Fluor 488 (clone N418, eBiosciences, San Diego; CA) or respective isotype controls for 1h at 4°C in the dark [26]. Each antibody was used in the presence of violet Live–Dead (1:10000; Life Technologies, Grand Island, NY) and 0.5 µg anti-CD16/CD32 blocking antibody (clone 2.4G2, BD Pharmingen, San Jose, CA). Samples were washed in 0.1% BSA/PBS, and fixed overnight in 4% formaldehyde in PBS (Thermo Scientific, Barrington, IL). On the next day cells were washed and suspended in 100 µl 0.1% BSA/PBS and six-hundred thousand events were collected on a LSRII flow cytometer (BD Biosciences, San Jose, CA). Data were analyzed in FlowJo 7.6.4 (Tree Star, Ashland, OR) and compensation was performed in FlowJo post collection. All analyses were performed on cells originally determined as vital Live-Dead negative cells. Migrated bmEos in the lungs were distinguished from alveolar macrophages (Siglec F^pos/CD11c^pos) as live Siglec F^pos/CD11c^neg cells and data were reported as percentage of vital cells. Migration in response to intratracheal hCCL24 was expressed as the in vivo chemotactic index (CI):

$$\mathrm{C}\mathrm{I}=\frac{\text{percentage of bmEos recoverd from hCCL}24-\text{treated mice}}{\text{percentage of bmEos recoverd from untreated control mice}}$$

In vitro chemotaxis assay

Mature bmEos were washed and resuspended at 10⁶ cells/ml in assay media (RPMI 1640/1% FCS/10 mM Hepes). Hundred µl of cell suspensions were placed into the upper wells, and 100 µl recombinant hCCL24 or vehicle controls were placed into the bottom wells of an equilibrated 96-well chemotaxis plate (Corning Life Sciences, Radnor, PA). Chemotaxis was stimulated for 3h at 37°C in a humidified CO₂ incubator. Migrated cells were enumerated for 30 s at high flow rate by flow cytometric counting on a FACSCalibur.

Migration in response to hCCL24 was expressed as the in vitro chemotactic index (CI):

$$\mathrm{C}\mathrm{I}=\frac{\#\text{cells migrated in response to hCCL}24}{\#\text{cells migrated in response to vehicle control}}$$

Statistical analysis

Data sets were analyzed using Student’s t-test unless otherwise indicated. All data are reported as the mean ± SEM.

Abbreviations

bmEos

bone marrow-derived eosinophils

GFP

green fluorescent protein

IL

interleukin

LT

leukotriene

PG

prostaglandin

hCCL24

human eotaxin-2

Tg

transgenic

WT

wild type