An essential role for Rab27a GTPase in eosinophil exocytosis

Essential role for the GTPase, Rab27a, in eosinophil exocytosis using in vitro and in vivo degranulation assays.

Abstract

Eosinophil degranulation has been implicated in inflammatory processes associated with allergic asthma. Rab27a, a Rab-related GTPase, is a regulatory intracellular signaling molecule expressed in human eosinophils. We postulated that Rab27a regulates eosinophil degranulation. We investigated the role of Rab27a in eosinophil degranulation within the context of airway inflammation. Rab27a expression and localization in eosinophils were investigated by using subcellular fractionation combined with Western blot analysis, and the results were confirmed by immunofluorescence analysis of Rab27a and the granule membrane marker CD63. To determine the function of eosinophil Rab27a, we used Ashen mice, a strain of Rab27a-deficient animals. Ashen eosinophils were tested for degranulation in response to PAF and calcium ionophore by measuring released EPX activity. Airway EPX release was also determined by intratracheal injection of eosinophils into mice lacking EPX. Rab27a immunoreactivity colocalized with eosinophil crystalloid granules, as determined by subcellular fractionation and immunofluorescence analysis. PAF induced eosinophil degranulation in correlation with redistribution of Rab27a⁺ structures, some of which colocalized with CD63⁺ crystalloid granules at the cell membrane. Eosinophils from mice had significantly reduced EPX release compared with normal WT eosinophils, both in vitro and in vivo. In mouse models, Ashen mice demonstrated reduced EPX release in BAL fluid. These findings suggest that Rab27a has a key role in eosinophil degranulation. Furthermore, these findings have implications for Rab27a-dependent eosinophil degranulation in airway inflammation.

Introduction

Eosinophils are inflammatory granulocytes implicated in asthma and parasitic helminth infections [1–4]. They are highly granulated white blood cells enriched in cationic granule proteins that are thought to contribute to tissue damage in inflammatory airway diseases, including atopic asthma [5]. Granule proteins from eosinophils are associated with phenotypes of asthma, which include, but are not limited to, hypersecretion of mucus, increased vascular permeability, bronchial epithelial damage, and smooth muscle contraction [4]. In humans, however, a proinflammatory role for eosinophils in allergic asthma remains controversial. Early studies using mepolizumab, a humanized anti-IL-5 monoclonal antibody, proposed a lack of efficacy in the asthma patients studied, thus concluding that eosinophils may not play a direct effector role in asthma [6–8]. However, 2 studies in 2009, using the same antibody therapy, indicated that eosinophil ablation significantly reduces the number of asthma exacerbations in patients with sputum eosinophilia [9, 10]. Another 2 studies in eosinophil-deficient mice have supported a proinflammatory role for eosinophils in asthma [11, 12]. On the basis of these studies and others, a role for eosinophils in airway hyperresponsiveness and remodeling in atopic asthma continues to be the focus of intensive research. Furthermore, the precise contribution of Rab27a to eosinophil effector functions, principally degranulation, has not been definitively established.

In the midst of this ongoing debate, Coughlin et al. [13] shed new light on a potential effector role for eosinophil Rab27a in asthma. Rab27a, a Rab-related GTPase, is involved in vesicular transport, budding, and movement, upstream of docking and fusion [14]. This secretory GTPase is usually membrane bound and becomes activated after displacement of bound GDP and association with GTP by the action of RabGEFs [15]. A rare, autosomal recessive disorder known as Griscelli syndrome type 2 is caused by a deficiency in Rab27a, characterized by partial albinism and immunodeficiency resulting in early childhood death, unless a bone marrow transplant is performed [16, 17]. This disorder is specifically associated with deficient exocytosis from immune and other secretory cells, including melanocytes. In our recent study, the expression and magnitude of activity of eosinophil Rab27a was compared between patients with asthma and control subjects. We reported significantly higher expression and activity of Rab27a in eosinophils from those with asthma [13]. From our data, we concluded that Rab27a and possibly eosinophil exocytosis are important contributors to an asthma phenotype.

The Ashen mouse strain was used to examine the in vivo relevance of this observation. The Ashen mouse is a Rab27a-deficient strain that produces a truncated Rab27a protein lacking critical GTP binding domains [18, 19]. These mice have diluted coat colors, related to reduced melanocyte secretion, and exhibit a profound immunodeficiency due to a failure to release cytolytic granules from CTLs and NK cells [20, 21]. Thus, many secretory cells from Ashen mice are characterized by abnormal exocytosis. We hypothesized that Rab27a deficiency reduces eosinophil exocytotic capacity. Eosinophils were isolated from Ashen mice and compared for their degranulation responses with IL-5 transgenic eosinophils. Ashen mice were also crossed with IL-5/hE2 double-transgenic mice to determine the role of Rab27a in in vivo eosinophil exocytosis. IL-5/hE2 double-transgenic mice are characterized by high expression of T-cell-derived IL-5, driven by the CD3 promoter, along with hE2 from Clara cells residing in the lungs, by fusion of the hE2 gene with the CC10 promoter [22]. These cytokines contribute significantly to eosinophil differentiation (IL-5) [23] and chemotaxis (hE2, including egress from bone marrow) [24]. IL-5/hE2 double-transgenic mice exhibit eosinophilia and spontaneously develop severe asthma-like symptoms with significant airway hyperresponsiveness. Thus, IL-5/hE2 double-transgenic mice serve as an abundant source of spontaneously degranulating eosinophils. In this study, IL-5/hE2 mice were crossed with the Ashen strain [18, 19], which then provided a Rab27a-deficient, eosinophil-producing model.

MATERIALS AND METHODS

Animals

All animals in this study were on a C57BL/6 background, and C57BL/6 WT mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). NJ.1638 IL-5 [23], hE2 [22], and IL-5/hE2 [22] transgenic mice were bred in house. These strains were crossed with Ashen [19] to generate IL-5/Ashen, hE2/Ashen, and IL-5/hE2/Ashen mice. The Ashen mice possessed similar numbers of total and differential peripheral blood eosinophils and leukocytes to that of WT C57BL/6 animals (Fig. 1). The IL-5/hE2 mice were hemizygous for both transgenes and served as controls for the IL-5/hE2/Ashen mice. Eosinophil peroxidase knockout (EPX^−/−) mice were generated that lacked EPX protein and activity [25]. (The acronym EPX is used here to avoid confusion with EPO, denoting erythropoietin.) IL-5/hE2 mice were also crossed with EPX^−/− mice to generate IL-5/hE2/EPX^−/− mice. The mice were maintained in ventilated microisolator cages housed in specific-pathogen–free environments at the University of Alberta (Edmonton, AB, Canada) and Mayo Clinic Scottsdale (Scottsdale, AZ, USA). All experiments were performed were in accordance with the guidelines of the University of Alberta's Animal Care and Use Committee, the Canadian Council of Animal Care, the National Institutes of Health, and the Mayo Foundation.

Figure 1. Total and differential peripheral blood leukocyte counts were similar between WT C57BL/6 and Ashen mice.

(A) Total cell counts from the peripheral blood of WT C57BL/6 and Ashen mice were similar. (B) Percentages of peripheral blood eosinophils were identical in WT and Ashen mice. (C) Differential counts of leukocytes in WT and Ashen mice showed similar proportions (n=3).

Preparation of eosinophils

HL-60 clone 15 (HL-60c15) cells (ATCC, Manassas VA, USA), chosen for this study based on their propensity to differentiate into highly granulated eosinophil-like cells, were treated with 500 μM n-butyrate [26]. Human eosinophils were purified from peripheral blood of atopic donors [27–29]. The purity and viability of human eosinophils was typically >98%. Splenic eosinophils were isolated from IL-5 transgenic mice by using a published method [30, 31]. For peripheral blood mouse eosinophils, the cells were recovered from the peripheral blood of IL-5 transgenic mice at >98% purity [23]. Cultured, in vitro bmEos were also generated from WT C57BL/6 mice and differentiated in the presence of recombinant mouse SCF, FLT3L, and IL-5 [32].

Subcellular fractionation and marker enzyme assays

Differentiated HL-60c15 cells and peripheral blood mouse eosinophils were homogenized in a cell cracker (a ball-bearing–containing cell homogenizer), and the resulting organelles were separated by a linear Nycodenz or Histodenz (Sigma-Aldrich St. Louis, MO, USA) gradient (0–45%) [33]. Marker enzyme assays were used on fractionated samples. EPX activity in the resulting fractions was measured by the reactivity with tetramethylbenzidine substrate, and cytosol fractions were detected with LDH or protein assay [33] (Bio-Rad, Hercules, CA, USA).

Immunofluorescence

Image analysis of the immunofluorescence of adherent eosinophils was performed after plating of peripheral blood mouse eosinophils from IL-5 transgenic NJ.1638 mice or human eosinophils from atopic donors in 8-well chamber slides (Lab-Tek Chambered Coverglass System; Nunc, Rochester, NY, USA) [34]. The cells were resuspended in serum-free RPMI medium, added to the glass surfaces of the chamber slides, and incubated for 15 min at 37°C, to allow adhesion. PAF (5 μM) was added to each well for 5 or 15 min. Stimulation was terminated by removal of the medium and fixation of the cells with 4% paraformaldehyde for 20 min. The cells were permeabilized with 0.1% Triton X-100 in PBS for 3 min before staining. The mouse eosinophils were then blocked for 20 min with 2% BSA+2% goat serum in PBS, to reduce nonspecific binding by antibodies, whereas the human eosinophils were blocked with 2% human IgG in PBS [35]. Mouse monoclonal anti-Rab27a (5 μg/ml; 20/RAB27 clone; BD Biosciences, Inc., San Jose, CA, USA) was then added and detected with anti-mouse IgG-Cy3, followed by fluorescein isothiocyanate (FITC)– mouse monoclonal anti-CD63 (Serotec, Raleigh, NC, USA). After the cells were washed with PBS, ProLong Gold mounting medium containing DAPI (Invitrogen-Molecular Probes, Eugene, OR, USA) was added to each well. Immunofluorescence labeling was imaged with a Deltavision OMX microscope equipped with a ×60 objective (1.43 NA; Applied Precision, Issaquah, WA, USA).

Gel electrophoresis and Western blot analysis

Proteins were separated by SDS-PAGE gels and transferred onto nitrocellulose membranes. The membranes were blocked with Odyssey blocking buffer (Li-Cor Bioscience, Lincoln, NE, USA) diluted 1:2 in PBS. Rab27a protein was probed with an anti-Rab27a rabbit antibody (Santa Cruz Biotechnology Inc., Dallas, TX, USA) [36]. Immunofluorescent protein bands were quantified by using IRDye-conjugated secondary antibodies against primary antibodies with a Li-Cor Odyssey infrared imaging system (Li-Cor Bioscience).

In vitro stimulation of mouse eosinophils

Eosinophils (2×10⁵) were stimulated with 0.1 μM PAF, 5 μM ionomycin, or both. The cells were initially centrifuged at 400 g for 10 min, and then supernatants from this step were centrifuged again at 10,000 g for 10 min at 4°C to clarify the supernatants for the assay. The resulting supernatants were placed in 96-well plates for detection of EPX by ELISA.

Preparation of BAL fluid

BAL fluid was recovered by introducing and collecting 1 ml of 2% FCS in PBS. BAL fluid was initially centrifuged at 400 g for 10 min at 4°C, after which the supernatant was recovered for a further centrifugation at 10,000 g for 10 min at 4°C to remove any remaining debris. Pellets from BAL samples were analyzed for total and differential cell counts.

Intratracheal instillation of mouse eosinophils

Peripheral blood eosinophils isolated from IL-5 transgenic mice were reconstituted at 10 × 10⁶ cells in 30 μl of PBS and adoptively transferred into the lungs of lightly anesthetized mice (3% isoflurane). BAL samples from these mice were collected 24 h after instillation of eosinophils.

ELISA of EPX

Quantification of EPX was achieved by an EPX ELISA method developed by Mayo Clinic Arizona [37]. EPX amounts were quantified by comparing them to the standard curve generated from measuring EPX in pure mouse eosinophil lysates. The results are shown as eos. equiv./microliter, to indicate the absorbance measured per eosinophil count.

Statistical analysis

Data were analyzed and graphed with Prism (version 5; GraphPad, San Diego, CA, USA). The results are represented as the mean ± se and analyzed by 1-way ANOVA with Tukey's post hoc analysis. For all experiments, P < 0.05 was considered to be statistically significant.

RESULTS

Rab27a was expressed in eosinophils and colocalized with EPX-containing granules

To determine whether mouse eosinophils express Rab27a, bmEos and peripheral blood mouse eosinophils were tested for expression of Rab27a mRNA and protein. We determined that Rab27a mRNA and protein were expressed in bmEos from WT C57BL/6 mice as well as bmEos and peripheral blood eosinophils from IL-5 transgenic NJ.1638 mice (Fig. 2). The levels of expression of Rab27a mRNA and protein, as well as the molecular weight of Rab27a protein, in mouse eosinophils was similar to that in human eosinophils and butyrate-treated HL-60c15 cells differentiated into an eosinophilic phenotype.

Figure 2. Rab27a was expressed in mouse eosinophils and colocalized with crystalloid granule fractions.

(A) RT-PCR analysis of differentiated HL-60c15 cells and bmEos cultured from C57BL/6 mice. (B) Western blot analysis of Rab27a expression in eosinophils from mouse and human sources. BmEos were isolated from C57BL/6 mice (left) and bmEos, and peripheral blood (PB) eosinophils were prepared from IL-5 Tg mice (right), to compare Rab27a expression with that in human eosinophils and differentiated HL-60c15 cells. A total of 20 μg protein was loaded in each lane. (C) Subcellular fractionation of peripheral blood eosinophils from IL-5 Tg mice. An enzymatic assay was used to detect EPX in fractions containing crystalloid granules (fractions 6–10), and LDH was assayed to identify cytosolic fractions (fractions 16–23). Western blot analysis showed Rab27a (fractions 6–13) and gp91^phox (a marker for plasma membrane, fractions 11–15).

To determine the subcellular localization of Rab27a, the eosinophils were subjected to lysis through a cell cracker, and the resulting intact organelles were separated across a linear density gradient consisting of Histodenz (0–45%). Peripheral blood eosinophils from IL-5 Tg mice expressed high EPX activity in crystalloid granule-containing fractions (high-density fractions 6–10) that substantially overlapped with fractions containing Rab27a (fractions 6–13) as determined by Western blot analysis. Most of the Rab27a immunoreactivity appeared in fractions enriched in membrane-bound organelles, which typically entered the gradient and was separate from cytosolic LDH-containing fractions above the gradient. The colocalization of Rab27a with membrane-bound organelles in fractionated eosinophils confirmed its membrane-associated properties, on the one hand [15]. Plasma membrane gp91^phox expression, on the other hand, corresponded to lower density, plasma membrane-containing fractions (fractions 11–15). These findings suggest that Rab27a colocalizes with EPX-containing crystalloid granules, as well as other lower density membrane-bound organelles, in unstimulated eosinophils.

Next, we determined the effects of PAF stimulation on the intracellular localization of Rab27a. PAF is a potent eosinophil secretagogue that induces EPX release [38]. Upon stimulation by PAF, both EPX activity and Rab27a expression shifted to a slightly higher density fraction (fraction 6) compared with that in the resting cells (Fig. 2C). The amount of total granule-associated EPX activity was also reduced in stimulated cells, suggesting degranulation of EPX. However, the distribution of Rab27a was not markedly altered after PAF stimulation and exhibited expression in both low- and high-density fractions (fractions 6–13) similar to that of unstimulated cells.

We confirmed colocalization of Rab27a with eosinophil crystalloid granules by performing immunofluorescence analysis of mouse and human eosinophils. CD63, used as a granule membrane marker for eosinophil crystalloid granules [39], showed that Rab27a colocalized with CD63⁺ granules in the cytoplasm of mouse and human eosinophils (Fig. 3A, B). CD63 is a membrane-bound tetraspanin molecule [39]. As expected, immunofluorescence for CD63 showed characteristic ring-like structures in human eosinophils, suggesting granule membrane localization of CD63. These intracellular CD63⁺ crystalloid granules were densely packed in human eosinophils and showed minimal plasma membrane localization. In mouse eosinophils, CD63⁺ ring structures were less prevalent, but similar to human eosinophils, CD63 was found only in intracellular structures resembling granules. In contrast, Rab27a was more widely distributed throughout the cells, with punctate staining indicating granular and vesicular localization, some of which colocalized with CD63⁺ crystalloid granules. These findings are in agreement with subcellular fractionation results.

Figure 3. Rab27a colocalized with crystalloid granules in mouse and human eosinophils.

(A) Mouse eosinophils from the peripheral blood of IL-5 Tg mice were plated onto glass coverslips in 8-chamber slides and allowed to adhere before fixation and staining for Rab27a and CD63, a membrane marker for crystalloid granules. Arrowhead indicates colocalization of Rab27a and CD63 to the membrane of a single crystalloid granule. (B) Human eosinophils (arrowhead) from peripheral blood were labeled similarly. (C) Mouse IL-5 Tg eosinophils were treated with PAF (5 μM) for 0, 5, and 15 min before fixation and staining. Arrowhead and inset show strong colocalization of Rab27a and CD63 at the cell periphery. (D) Human peripheral blood eosinophils also demonstrated cell-stretching, along with granule polarization at the leading edge, with colocalization of Rab27a and CD63 at the edges of the cells during stimulation. Images of stimulated cells are representative of 30–40% of cells in 10 high-powered fields each. Scale bar = 5 μm.

Upon stimulation with PAF, several CD63⁺ granules redistributed toward the periphery and plasma membrane in both mouse (Fig. 3C) and human (Fig. 3D) eosinophils. Granule polarization to leading edges in both human and mouse eosinophils was evident within 5 min of stimulation with PAF, before EPX release. Previous studies showed that PAF at 5 μM induces significant EPX release after 30 min of stimulation [38]. PAF was applied to cell supernatants and not as a chemotactic gradient, suggesting that polarization of granules occurred within the cells and that degranulation from PAF-stimulated eosinophils involves polarized exocytosis.

In parallel with that finding, immunofluorescence for Rab27a showed colocalization with CD63⁺ granules in unstimulated cells (Fig. 3). However, consistent with subcellular fractionation data, Rab27a was not exclusively localized to CD63⁺ granules in mouse and human eosinophils, as other Rab27a⁺ structures were evident that did not colocalize with CD63. Upon stimulation of mouse eosinophils with PAF (5 min), Rab27a appeared to redistribute away from the CD63⁺ granules found within cytoplasmic sites to other locations in the cell, including the cell membrane (Fig. 3C). This selective redistribution of Rab27a⁺ organelles is suggestive of piecemeal degranulation, in which small, rapidly mobilized secretory vesicles shuttle granule products from the crystalloid granules to the cell membrane, as previously shown in human eosinophils [29]. However, at 15 min of PAF stimulation, Rab27a and CD63 colocalized to highly focused, punctate regions at or near the cell membrane in mouse eosinophils. In human eosinophils, the pattern of Rab27a redistribution was distinct. PAF induced a redistribution of Rab27a that colocalized with CD63⁺ structures in polarized regions of cells, with relatively little concentration at the cell membrane (Fig. 3D, arrowhead).

Ashen eosinophils showed deficient EPX release compared with WT eosinophils in vitro

Rab27a is a critical regulatory protein in granule exocytosis in other secretory cells, and so we examined the role of Rab27a in degranulation from eosinophils. We isolated eosinophils from Ashen mice, a substrain of C57BL/6 mice that contains a splicing mutation in Rab27a [18], that were crossed with IL-5-overexpressing mice (NJ.1638) [23], to obtain a large number of splenic and peripheral blood eosinophils lacking Rab27a expression. We then compared degranulation responses of eosinophils from the resulting IL-5/Ashen offspring with those of the respective WT controls against the same background strain (C57BL/6) as was crossed with IL-5 transgenic mice.

Peripheral blood eosinophils from IL-5/WT and IL-5/Ashen mice were stimulated with the secretagogues PAF, ionomycin, or PAF+ionomycin, with EPX release being quantified as eos. equiv./microliter. This unit of measurement was used to ensure that peripheral eosinophils isolated from both types of mice contained equal amounts of EPX, as measured by an in-house ELISA [37]. Using this approach, we were able to detect increased EPX release in supernatants from eosinophils stimulated with PAF, the calcium ionophore ionomycin, and a combination of PAF and ionomycin (Fig. 4). In contrast, IL-5/Ashen exhibited significantly reduced EPX release in response to all stimuli, compared with that in IL-5/WT eosinophils (Fig. 4).

Figure 4. Eosinophils from IL-5/Ashen mice were deficient in EPX release in vitro.

In vitro degranulation was observed in purified peripheral blood eosinophils from IL-5 Tg and IL-5/Ashen mice. Supernatants from stimulated cells were measured by EPX ELISA. Cells were stimulated with 50 ng/ml PAF and/or 1 μM ionomycin for 6 h. The data are the mean ± se (n≥10). *P < 0.05, ***P < 0.001.

Ashen eosinophils exhibited defective EPX release in vivo

We next confirmed that Ashen eosinophils are defective in their degranulation responses when administered in vivo. WT and Ashen mice were sensitized and challenged with OVA, and EPX release was determined by ELISA of BAL samples obtained on day 28 (Fig. 5). WT and Ashen mice treated with saline alone did not show any difference in EPX levels in their BAL. However, OVA sensitization and challenge of Ashen mice led to a significant (P<0.001) decrease in EPX content in BAL samples compared with those from OVA-treated WT mice. These findings suggest that airway eosinophil degranulation is deficient after OVA sensitization and challenge in Ashen mice.

Figure 5. OVA-sensitized and -challenged Ashen mice exhibited decreased BAL EPX release.

(A) EPX was detected in BAL samples obtained on day 28 of OVA sensitization and challenge or saline controls from WT and Ashen mice. EPX quantity was measured by EPX ELISA. (B) BAL cellularity, (C) number of BAL eosinophils, and (D) BAL percentage of eosinophils in saline- and OVA-challenged WT and Ashen mice. The data are the mean ± se (n≥5). *P < 0.05; **P < 0.01; ***P < 0.001.

However, upon further assessment of total BAL cellularity, we observed a reduced total number of cells and eosinophils in the BAL of OVA-treated Ashen compared with that of similarly treated WT mice, even though the percentages of eosinophils of total BAL cellularity were similar in both (Fig. 5B–D). Thus, reduced EPX concentrations in BAL samples from Ashen mice may be related to a decrease in eosinophils rather than a degranulation defect.

The question of whether the reduced BAL EPX quantities observed in OVA-treated Ashen mice was due to lower eosinophil counts was addressed in a novel ex vivo eosinophil degranulation experiment. Peripheral blood eosinophils from IL-5/WT and IL-5/Ashen mice were isolated, purified, and adoptively transferred into the lungs of triple-transgenic IL-5/hE2/EPX^−/− mice. These mice were generated to evoke maximally stimulating eosinophil degranulation conditions in vivo, with a complete absence of endogenous EPX secreted by the recipient animal. Typically, a large number of eosinophils transmigrate to the airway spaces in IL-5/hE2 double-transgenic animals and spontaneously degranulate in response to tissue-overexpressed IL-5 and eotaxin-2 [22]. Thus, intratracheally introduced WT eosinophils would be expected to degranulate in the airways of these animals (due to transgenic overexpression of eosinophil-specific activating cytokine IL-5 and chemokine eotaxin-2). In this ex vivo degranulation assay system, any EPX that is detected in BAL samples would be exclusively derived from donor eosinophils instilled into the trachea of IL-5/hE2/EPX^−/− mice.

Thus, we compared EPX levels from BAL supernatants of mice with intratracheal instillation of IL-5/WT and IL-5/Ashen peripheral blood eosinophils (Fig. 6). BAL from mice that received IL-5/WT eosinophils exhibited robust EPX release. In contrast, BAL from IL-5/Ashen eosinophil-instilled mice showed significantly reduced amounts of EPX (P<0.001) (Fig. 6A). When differential cell counts were performed on these samples, no differences were observed in total BAL cellularity and eosinophil count (Fig. 6B, C). Thus, these results support and confirm our data from EPX levels in BAL of OVA-treated mice and suggest that Rab27a-deficient Ashen eosinophils have defective degranulation responses in vivo.

Figure 6. IL-5/Ashen eosinophils were defective in EPX release when injected intratracheally.

(A) Purified peripheral blood eosinophils (10⁷) from IL-5/WT or IL-5/Ashen mice were intratracheally instilled into IL-5/hE2/EPX^−/− mice, and BAL was collected from euthanized mice 24 h after instillation. (B) BAL cellularity and (C) BAL eosinophils after intratracheal injection of eosinophils into IL-5/hE2/EPX^−/− mice. The data are the mean ± se (n=4). ***P < 0.001.

We tested the apparent degranulation deficiency of Ashen eosinophils further in IL-5/hE2/WT and IL-5/hE2/Ashen mice. These were generated by crossing double-transgenic IL-5/hE2 mice with either WT C57BL/6 or Ashen mice and using the offspring that were positive for the appropriate genotypes. For comparison, a third strain of IL-5/hE2/EPX^−/− mice was generated in which EPX expression was absent. BAL supernatants from these 3 strains of mice (6–8 wk) were collected, and spontaneously secreted EPX levels were measured (Fig. 7). Again, IL-5/hE2/Ashen mice showed markedly reduced EPX levels in BAL than in IL-5/hE2/WT BAL. Taken together, our data demonstrate that eosinophils lacking Rab27a have a deficiency in secretion of EPX-containing crystalloid granules.

Figure 7. Spontaneous release of EPX was reduced in BAL fluid from IL-5/hE2/Ashen mice.

IL-5/hE2/Ashen mice (6–8 wk) were generated to determine spontaneous EPX release in BAL fluid in comparison with that in age-matched IL-5/hE2 mice. EPX was measured by EPX ELISA. The data are the mean ± se (n=3–4). ***P < 0.001.

DISCUSSION

This report describes, for the first time, a direct role for Rab27a in eosinophil degranulation, identified in the Ashen Rab27a gene knockout strain. The role of Rab27a in granule exocytosis and the development of inflammatory processes has been investigated in CTLs [20, 21, 40], NK cells [16, 41], and neutrophils [42], whereas Rab27b has been implicated in mast cells [43]. Our previous report on the expression and function of Rab27a in human eosinophils demonstrated that eosinophils express and activate Rab27a in response to artificial agonists (phorbol esters and calcium ionophore) as well as physiological stimuli (fMLF and serum-coated zymosan) [13]. Human peripheral blood eosinophils expressed mRNA for Rab8a, -8b, -10, -11a, -27a, and -37, with weak signals for Rab11b and -13 and no detectable signal for Rab27b. We also found increased expression of Rab27a protein, with greater activity observed in eosinophils from patients with asthma compared with normal subjects [13]. However, a direct role had not been shown for Rab27a in exocytosis of crystalloid granules in eosinophils.

The function of Rab27a in granule exocytosis is dependent on a wide range of specific effector molecules [44]. Proteins containing an N-terminal Rab27a-binding region, known as Slp/Slac2 proteins or exophilins, have been shown to bind Rab27a and mediate respiratory burst in neutrophils [42, 45]. Rab27a may also bind directly to SNARE fusion proteins, to initiate exocytotic membrane fusion between granules and the plasma membrane. The R-SNAREs VAMP-2, -7, and -8, along with their respective binding partners, the Q-SNAREs SNAP-23 and syntaxin-4, are expressed in human eosinophils, and VAMP-7 has been shown to regulate exocytotic release of the granule proteins EPX and EDN [46–49]. Finally, another putative effector molecule for Rab27a may be the Sec/Munc family member Munc13-4, which has been demonstrated in other cell types, including CTL cells, NK cells, mast cells, and platelets, to regulate Rab27a-induced exocytosis [50–52].

The expression and function of Rab27a and its effector molecules have not yet been explored in eosinophils. These observations led to the current study with a view to further elucidate the expression and function of Rab27a in eosinophils and to determine how this protein may influence degranulation. Using Western blot analysis, subcellular fractionation, and immunofluorescence analysis, we established that Rab27a was expressed in both mouse and human eosinophils and that it colocalized with eosinophil crystalloid granules, as well as other membrane-bound organelles. To pursue this finding, we chose the Ashen strain, which exhibits a functional deficiency in Rab27a, and derived mouse bmEos from those animals, to confirm the function of Rab27a in mouse eosinophils and validate ensuing mouse models of airway inflammation. Previous studies in neutrophils have shown that Rab27a is mostly localized to specific and gelatinase-enriched tertiary granules and is not essential for myeloperoxidase release from azurophilic granules [53]. In contrast, our results provide strong evidence that Rab27a associates with eosinophil crystalloid granules containing CD63 and EPX and that Rab27a is necessary for EPX release.

To determine whether the localization of Rab27a and EPX change with stimulation, eosinophils were treated with the potent secretagogue PAF [38]. After PAF activation, the total EPX level across subfractions was significantly reduced, suggesting extracellular release of EPX, and EPX activity was split between 2 fractions (6 and 8). A possible explanation for the double peak of EPX activity in stimulated eosinophils is that stimulated cells release EPX through piecemeal degranulation [29, 54]. In confirmation of this finding, immunofluorescence images showed that a proportion of Rab27a redistributed away from CD63⁺ granules after 5 min of PAF stimulation in mouse eosinophils. Some Rab27a immunoreactivity was found at the cell membrane, colocalizing with a proportion of CD63⁺ granules in mouse eosinophils. However, at 15 min of PAF stimulation, there was a redistribution of Rab27a⁺ structures colocalizing with CD63⁺ granules at the cell membrane in highly focused, punctate regions. This suggests that Rab27a may be involved in piecemeal degranulation in early cell stimulation, as well as crystalloid granule fusion with the cell membrane in mouse eosinophils at later stages of degranulation (within 15 min of PAF stimulation).

In human eosinophils, PAF induced a marked polarization, whereupon CD63⁺ granules mobilized to the leading edges of the cells. Rab27a was found to colocalize with some CD63⁺ granules at the leading edge of the cells. Taken together, these findings suggest that Rab27a coordinates fusion between crystalloid granules and plasma membrane to regulate granule exocytosis in eosinophils, although there were subtle differences in the manner in which Rab27a was redistributed after PAF stimulation in mouse and human cells.

The role of Rab27a in eosinophil degranulation was further investigated by stimulating peripheral blood eosinophils isolated from WT and Ashen mice. In this system, a novel EPX ELISA [37] was used to measure the amount of released mouse EPX after stimulation. This EPX ELISA was developed in house and provides a 10-fold higher sensitivity than does the widely used OPD assay [55]. Some studies have suggested that mouse eosinophils do not readily degranulate either in vitro or in vivo [13, 56, 57]. However, we have recently reported that mouse eosinophils degranulate in vitro in response to at least 2 secretagogues: PAF and calcium ionophore [34, 38].

Our in vivo findings confirmed that Rab27a plays an important role in eosinophil exocytosis in association with EPX-containing crystalloid granules. OVA-treated Ashen BAL samples exhibited significantly diminished levels of EPX when compared to OVA-treated WT BAL. However, when differential cell counts were performed, OVA-treated Ashen mice had lower BAL cellularity, despite having a percentage of eosinophils similar to that in OVA-treated WT mice. This result suggests that the reduction of EPX observed in the BAL of OVA-treated Ashen mice is related to a reduction in airway eosinophils. The finding was unexpected, but was not surprising, given that global gene deletion of Rab27a may indirectly affect eosinophil recruitment and accumulation in the lungs through cytokine and chemokine secretion from other cell types that are dependent on Rab27a.

To further explore the possibility that Rab27a is involved in eosinophil degranulation, a novel ex vivo degranulation assay for eosinophils was designed in which eosinophils were intratracheally injected into IL-5/hE2/EPX^−/− mice. The use of IL-5/hE2/EPX^−/− mice provided a unique lung environment where intratracheally instilled eosinophils are induced to degranulate by in vivo secretagogues generated from tissue sources. These mice lack systemically expressed EPX, and any detected EPX in BAL samples must therefore originate exclusively from donor, not recipient, eosinophils. We found significantly less EPX in BAL from IL-5/hE2/EPX^−/− mice with IL-5/Ashen eosinophils instilled when compared with those with IL-5/WT eosinophils instilled, and differential cell counts performed on BAL from these 2 groups were not significantly different. These data provide supportive evidence for Rab27a in eosinophil degranulation, using a model in which exogenously administered eosinophils are introduced into a highly stimulatory airway environment. Consistent with these findings, double-transgenic IL-5/hE2 mice crossed with Ashen mice to generate a large number of eosinophils lacking Rab27a also showed a significant decrease in EPX in BAL samples from IL-5/hE2/Ashen mice.

In conclusion, our findings provide strong evidence supporting a role for Rab27a in eosinophil degranulation, both in vitro and in vivo, in the mouse model. The findings support our study on Rab27a function in human eosinophils [13] and suggest that eosinophils are dependent on Rab27a for degranulation responses in allergic airway inflammation and asthma.

AUTHORSHIP

J.D.K., L.W., N.S., R.H., and A.S. conducted the experiments and were supervised by J.J.L., R.M., and P.L. P.L. also conducted experiments for the study. S.O. contributed to experiments and training of users and was supervised by J.J.L. M.C.S. provided the Ashen strain. This study was supervised by J.J.L., R.M., and P.L., who obtained the funding.

DISCLOSURES

The authors declare no conflicts of interest.