Peripheral blood eosinophils from patients with allergic asthma contain increased intracellular eosinophil-derived neurotoxin

Peripheral blood eosinophils from patients with allergic asthma contain increased intracellular eosinophil-derived neurotoxin Julie B. Sedgwick, PhD,a Rose F. Vrtis, BS,a Kristyn J. Jansen, BS,a Hirohito Kita, MD,b Kathleen Bartemes, BA,b and William W. Busse, MDa Madison, Wis, and Rochester, Minn Mechanisms of asthma and allergic inflammation

Background: One mechanism of the eosinophil’s contribution to airway inflammation in asthma is through release of cationic granule proteins to cause airway injury. Differences in either the intracellular concentration of granule proteins or the extent of activated degranulation between eosinophils from healthy patients and those with allergy and asthma could, therefore, relate to fundamental differences in this cell’s function. Objective: To identify phenotypic differences in eosinophilderived neurotoxin (EDN) content and release in eosinophils from healthy patients, those with allergy, and those with allergy and asthma. Methods: Peripheral blood eosinophils were isolated by negative anti-CD16 selection. Total intracellular and cytokineactivated release of EDN protein was measured by radioimmunoassay. EDN mRNA was assessed by real-time PCR. Results: Eosinophils from patients with asthma contained significantly more EDN per cell than comparable cells from healthy patients, those with allergy but without asthma, or those with asthma treated with inhaled corticosteroids, but they had concentrations similar to airway eosinophils isolated from bronchoalveolar lavage fluid 48 hours after segmental bronchoprovocation with allergen. Furthermore, this increased granule protein was reflected in more EDN degranulation by IL-5– or GM-CSF–activated eosinophils when calculated as nanograms of protein secreted but not when calculated as a percentage of total EDN release. Levels of EDN mRNA were similar in all subject groups. Conclusions: These data suggest that peripheral blood eosinophils from subjects with untreated asthma have increased inflammatory capacity, as reflected by greater intracellular concentrations of EDN. (J Allergy Clin Immunol 2004;114:568-74.) Key words: Eosinophil, eosinophil-derived neurotoxin, degranulation, asthma, allergic rhinitis, corticosteroids, hypereosinophilic syndrome

The relationship between allergic rhinitis and asthma has been extensively studied, yet a final consensus has not emerged as to specific differences other than altered airway function in these diseases. The concept of ‘‘one From aAllergy and Immunology, Department of Medicine, University of Wisconsin, Madison; and bthe Mayo Clinic, Rochester. Received for publication September 19, 2003; revised May 11, 2004; accepted for publication May 12, 2004. Reprint requests: Julie Sedgwick, PhD, Department of Medicine, CSC-3244, H6/355 Allergy, 600 Highland Ave, Madison, WI 53792. E-mail: [email protected]. 0091-6749/$30.00 Ó 2004 American Academy of Allergy, Asthma and Immunology doi:10.1016/j.jaci.2004.05.023

568

Abbreviations used AA: Patients with mild allergic asthma AA+ICS: AA subjects treated with ICS AR: Subjects with allergic rhinitis BAL: Bronchoalveolar lavage ECP: Eosinophil cationic protein; ribonuclease 3 EDN: Eosinophil-derived neurotoxin; ribonuclease 2 FMLP: N-formyl-methionyl-leucyl-phenylalanine HBSS/gel: HBSS + 0.03% gelatin HES: Hypereosinophilic ICS: Inhaled corticosteroid LTF: Lactoferrin MPO: Myeloperoxidase NCS: Newborn calf serum NL: Healthy subjects without allergy or asthma rRNA: Ribosomal RNA SBP-AG: Segmental bronchoprovocation with allergen

airway, one disease’’ has been proposed by a World Health Organization–based position statement on the pathophysiological, epidemiologic, clinical, and treatment similarities between allergic rhinitis and allergic asthma.1,2 In this model, allergic diseases make up a continuum of severity from mild allergic rhinitis of the upper respiratory tract to allergic asthma with increased airway obstruction, bronchial hyperresponsiveness, and chronic inflammation of the lower respiratory tract. Inflammation of both the upper and lower respiratory tracts includes infiltration of eosinophils after allergen exposure.3 In chronic airway inflammation in asthma, eosinophils can release multiple inflammatory mediators, including granule proteins, reactive oxygen species, arachidonic acid metabolites, cytokines, and chemokines. Because eosinophils have been strongly implicated in the pathophysiology of asthma, it is reasonable to hypothesize that these cells are functionally different in allergic nasal disease and asthma. There have been, however, no reports as to distinct differences in eosinophil function between these diseases. Morphologically, the eosinophil is characterized by its content of cationic granules, whose release has been implicated in many manifestations of asthma, including inflammation, bronchial hyperresponsiveness, and airway obstruction.4 Recent studies have determined serum eosinophil granule protein concentrations as a predictor of asthma severity or activity. Increased serum levels of eosinophil granule proteins are reported for adults and

Sedgwick et al 569

J ALLERGY CLIN IMMUNOL VOLUME 114, NUMBER 3

METHODS Reagents and cytokines Percoll was purchased from Pharmacia (Uppsala, Sweden). HBSS, RPMI-1640, PBS, newborn calf serum (NCS), FCS, and trypan blue were obtained from Life Technologies (Grand Island, NY). Other reagents were purchased from Sigma Chemical Co (St Louis, Mo) unless otherwise stated.

Human subjects Healthy subjects without allergy or asthma (NL) were skin prick test negative to inhalant allergens, had no history of allergies or other underlying medical conditions, and were taking no medications. Subjects with allergic rhinitis (AR) had a history of seasonal or perennial allergic rhinitis and were skin prick test positive (>3 mm by skin prick test) to at least 1 of 12 common allergens tested, including cat, ragweed, and house dust mite; they had no clinical history of asthma. These subjects’ only medications were antihistamines or inhaled nasal corticosteroids as needed. Patients with mild allergic asthma (AA) were skin test positive, hyperresponsive on methacholine challenge (PC20 < 8 mg/mL) and had a diagnosis of clinical asthma. These subjects used only inhaled b2 antagonists and had not been treated with oral or inhaled corticosteroids (ICS) within 6 months of participation in this study. AA subjects treated with ICS for more than a year were designated a separate group (AA+ICS). Finally, peripheral eosinophils from a nonsymptomatic untreated HES patient were isolated. Informed, written consent was obtained from all participants before inclusion in the study, which had been approved by the University of Wisconsin Institutional Review Board.

Blood eosinophil isolation Eosinophils were isolated by using negative immunomagnetic bead selection as previously described.13 Eosinophils were >97% pure and >98% viable; contaminating cells were neutrophils and lymphocytes.

Airway eosinophil isolation Seven additional AA subjects underwent segmental bronchoprovocation with allergen (SBP-AG) (ragweed, cat, or house dust mite) followed at 48 hours by bronchoalveolar lavage (BAL).14 The BAL cells were washed and layered over multidensity gradients that consisted of 1.085 and 1.100 g/mL Percoll. After centrifugation, the cell band at the 1.085/1.100 interface was collected and consisted of >97% BAL eosinophils with >98% viability by trypan blue dye exclusion. The contaminating cells were neutrophils and lymphocytes. Blood was drawn from these subjects just before the 48-hour bronchoscopy and was processed as described previously to isolate peripheral blood eosinophils.

Blood neutrophil isolation Neutrophils were isolated from peripheral blood by dextran sedimentation of red blood cells and density fractionation of the buffy coat over Percoll 1.077 g/mL. The cell pellets were collected, contaminating red blood cells were lysed, and the resulting neutrophils were >95% pure with >98% viability. Contaminating cells were eosinophils.

Total intracellular and released eosinophil EDN Total EDN was determined by incubating eosinophils (2 3 106/mL HBSS + 0.03% gelatin [HBSS/gel]) with an equal volume of 1% Triton X-100 in HCl 0.1 mol/L. After a 4-hour incubation, the cellfree supernates were collected and stored at ÿ208C until assayed for EDN by RIA (threshold of detection, 2 ng/mL).15 Similar samples were analyzed for intracellular ECP by UniCap ECP (Pharmacia Diagnostics, Kalamazoo, Mich) according to the manufacturer’s directions (threshold of detection, 2 ng/mL). For degranulation experiments, eosinophils (1 3 106/mL) were activated with HBSS/ gel, 100 nmol/L N-formyl-methionyl-leucyl-phenylalanine (FMLP), IL-5 10 ng/mL, or GM-CSF 10 ng/mL for 4 hours at 378C in a 5% CO2 humidified incubator. The cell-free supernates were collected and stored at ÿ208C until assayed for EDN.

Total cellular and released neutrophil total cellular lactoferrin and myeloperoxidase Total cellular lactoferrin (LTF) and myeloperoxidase (MPO) were determined by incubating neutrophils (2 3 105/mL HBSS/gel) with an equal volume of 4% Triton X-100 for 15 minutes. LTF and MPO levels were measured in the samples by commercial ELISA according to the manufacturer’s directions (Oxis Research, Portland, Ore). Neutrophil degranulation was measured by incubating 1 3 105 cells per milliliter with selected agonists in the presence of cytochalasin B 5 lg/mL for 15 minutes at 378C.16,17 The cell-free supernates were collected and stored at ÿ208C until assessed by ELISA for LTF and MPO release (thresholds of detection, 1.5 ng/mL).

Semiquantitative real-time PCR for detection of EDN mRNA Total RNA was extracted from 4.5 3 106 blood eosinophils by using a 1-step phenol/chloroform extraction reagent (Tri Reagent) according to the manufacturer’s instructions. The total RNA was treated with deoxyribonuclease (RQ1 ribonuclease-free deoxyribonuclease, Promega, Madison, Wis) to degrade genomic DNA and was then resuspended in 35 lL of Nuclease-Free Water (Promega).

Mechanisms of asthma and allergic inflammation

children with asthma compared with various subjects without asthma.5-7 Matsumoto et al8 found serum eosinophil cationic protein (ECP; ribonuclease 3) levels to be associated with subject age and disease duration in adult asthma, whereas others observed that the total intracellular granule protein content of eosinophils from healthy subjects, patients with allergy, and hypereosinophilic (HES) subjects was related to disease activity and to the density and concentration of circulating eosinophils.9-12 These studies have led to disparate results that are due to different granule proteins studied, assays used, and patient populations compared. To our knowledge, a comparison of the relative concentrations of a single intracellular granule protein in eosinophils from healthy subjects and those with allergy and asthma has not been reported. Because phenotypic differences in eosinophils in patients with asthma may either lead to or be associated with greater airway injury, we hypothesized that the total intracellular content of these cationic proteins may be dependent on disease status. Moreover, we postulated that eosinophils compartmentalized to the circulation versus those in the airway contain distinct amounts of granule proteins because of differential in vivo exposure to inflammatory mediators. To test this hypothesis, peripheral blood eosinophils were isolated from healthy nonatopic volunteers without asthma or from patients with allergic rhinitis or allergic asthma to determine total cellular eosinophil-derived neurotoxin (EDN; ribonuclease 2) content and its release upon degranulation.

570 Sedgwick et al

Mechanisms of asthma and allergic inflammation

Reverse transcription was performed on 16 lL of the total RNA solution with 1 lg of random hexamer primers (Promega) for 2 minutes at 708C, followed by the addition of 400 U of reverse transcriptase (Superscript II RT, Invitrogen, Carlsbad, Calif ), 8 lL of 53 reaction buffer, 4 lL of dithiothreitol 0.01 mol/L (Invitrogen), 80 U of recombinant RNasin (Promega), and 0.02 lmol of deoxynucleoside triphosphates (Promega), in a total volume of 42 lL, with an incubation of 378C for 1 hour and then 948C for 5 minutes. Real-time PCR was used to assay for EDN mRNA and ribosomal RNA (rRNA) and was performed on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, Calif ). Duplicate 25-lL samples were analyzed in a 96-well plate with an optical cover. The reaction mixture included 13 TaqMan Universal PCR Master Mix (Applied Biosystems) plus primers (900 nmol/L) and probe (200 nmol/L), along with either 2 or 5 lL of cDNA test sample to measure rRNA or EDN mRNA, respectively. The EDN forward primer was 59-GGATCAGTTCTCACAGGAGCTACA-39; the reverse primer was 59-CCCAACAGAGAGAGCAGACAAA-39. The EDN probe 59-CTGGGAAACATGGTTCCA-39 was designed to span an intron and was labeled with 6-carboxyfluorescein as the reporter with a nonfluorescent quencher. The EDN primers and probe were designed from NCBI GenBank Accession File No. M28129 by using Primer Express software and were manufactured by Applied Biosystems. The rRNA was measured with a commercial TaqMan Ribosomal RNA Control Reagent (catalog No. 4308329, Applied Biosystems). The thermal cycler protocol was 508C for 2 minutes, 958C for 10 minutes, 958C for 15 seconds, and 608C for 1 minute for 40 cycles. A standard curve using a strongly positive sample was included in each assay to calculate a relative quantity and to check the assay for linearity. A cycle threshold was selected in the geometric phase of amplification. Results are expressed as relative mRNA units.

Statistics Data are presented as means ± SEMs, and the groups were analyzed by 1-way ANOVA with repeated measures. When the ANOVA demonstrated significant differences between data groups, it was followed by pairwise comparisons with the Tukey test for parametric data or the Dunn test for nonparametric data (SigmaStat, SSPS, Chicago, Ill). A P value < 0.05 was considered significant.

RESULTS Subject characteristics Subjects ranged in age from 18 to 60 years and had an equal sex distribution (Table I). NL peripheral blood eosinophil counts were significantly lower than in the other subject groups. Although the NL, AR, AA, and BAL groups had similar values for %FEV1 and % forced vital capacity, the AA+ICS subjects had significantly lower FEV1 and forced vital capacity values than the AR and BAL subjects. Total intracellular EDN Eosinophils from NL and AR subjects had similar concentrations of total cellular EDN, whereas cells from AA subjects were significantly more numerous (Fig 1). It is interesting to note that eosinophils from AA subjects treated with ICS (>1 year) had EDN concentrations similar to NL and AR subjects but lower concentrations than in untreated AA subjects. Eosinophils isolated from the peripheral blood and BAL fluid of a group of AA subjects 48 hours after they had undergone SBP-AG had similar

J ALLERGY CLIN IMMUNOL SEPTEMBER 2004

EDN levels, and these were not significantly different from EDN concentrations found in the other groups. Eosinophils from the single HES subject had the lowest level of intracellular EDN. Although there was a trend toward decreased intracellular ECP in eosinophils from AA subjects with or without ICS treatment (NL: 2172 ± 116 ng/1 3 106 eosinophils, n = 5; AR: 1932 ± 298 ng/1 3 106 eosinophils, n = 12; AA: 1827 ± 180 ng/1 3 106 eosinophils, n = 15; AA+ICS: 1794 ± 179 ng/1 3 106 eosinophils, n = 5), there were no significant differences in ECP among groups. No significant differences were found in plasma EDN from the NL (17.4 ± 2.9 ng/mL), AR (20.7 ± 3.2 ng/mL), AA (28.6 ± 7.8 ng/mL), or AA+ICS (31.2 ± 8.4 ng/mL) groups, even though a trend was found toward higher plasma EDN levels with increasing disease severity. Intracellular eosinophil EDN concentrations did not correlate with plasma levels for any of the subject groups (data not shown).

Eosinophil degranulation Eosinophils from the same study groups were incubated with IL-5 (10 ng/mL), GM-CSF (10 ng/mL), or FMLP (100 nmol/L) to activate degranulation. When degranulation was calculated as nanograms of EDN released per 1 3 106 cells, circulating eosinophils from AA subjects released significantly more EDN than from AR subjects after activation with IL-5 or GM-CSF (Fig 2, A). BAL eosinophils from AA subjects, in contrast, released significantly less EDN in response to IL-5 and GM-CSF compared with all other groups. No differences among groups were found with FMLP stimulation or spontaneous degranulation. In contrast, peripheral blood eosinophils from all of the groups demonstrated equivalent EDN release when degranulation was calculated as a percentage of the total cellular EDN (Fig 2, B). Again, BAL eosinophils released a much lower percentage of their cellular EDN when activated by IL-5 (P < .05 vs all groups) or GM-CSF (P < .05 vs NL and AA groups). EDN mRNA To determine whether the increased intracellular EDN was a result of or was reflected in altered EDN transcription, EDN mRNA was measured by real-time PCR and normalized against rRNA. The mean EDN mRNA/rRNA ratios were equivalent for all groups (range, 2-3.2). In contrast, mRNA/rRNA was substantially higher (14.9) for the HES subject. No correlations were observed between cellular EDN concentrations and mRNA levels. Neutrophil granule proteins To determine whether the increased concentration of EDN protein in the AA eosinophils was reflected in neutrophil granule protein content, total cellular LTF and MPO were measured in the same subjects. These neutrophil enzymes were similar in all subject groups (LTF [ng/1 3 105 neutrophils]: NL, 198 ± 28; AR, 159 ± 31; AA, 186 ± 26; AA+ICS, 207 ± 36; MPO: [ng/1 3 105

Sedgwick et al 571

J ALLERGY CLIN IMMUNOL VOLUME 114, NUMBER 3

TABLE I. Patient characteristics

n Mean age, y (range) Blood eosinophils (/mm3) % FEV1  % FVC  No. Positive skin prick tests

Normal (NL)

39 118 99 106.5

7 (25-53) ± 16* ± 6.1 ± 8 0

Allergic asthma (AA)

10 39 (25-60) 208 ± 29 111.8 ± 4.4 119.4 ± 5§ 1-4

11 36 (22-60) 244 ± 43 98.2 ± 3.9 100.7 ± 3.6 1-12

AA+ICS

38 260 79 90.3

5 (24-50) ± 73 ± 2.3à ± 4.9 1-3

HES

1 24 4680

AA 48 h after SBP-AG (BAL and blood)

29 322 101.4 102.8

7 (20-48) ± 50 ± 3.5 ± 3.7 $1 Mechanisms of asthma and allergic inflammation

Variable

Allergic rhinitis (AR)

FVC, Forced vital capacity. *P < .05 versus AR, AA, AA+ICS, and AA after SBP-AG.  Percentage of predicted. àP < .05 versus AR and AA after SBP-AG. §P < .05 versus AA and AA+ICS.

neutrophils]: NL, 300 ± 78; AR, 310 ± 37; AA, 267 ± 90; AA+ICS, 311 ± 10; n = 3-7). Moreover, no significant differences were observed in neutrophil LTF or MPO degranulation after activation with FMLP 100 nmol/L, platelet-activating factor 100 nmol/L, or IL-8 1 to 100 ng/ mL (data not shown). GM-CSF did not stimulate neutrophil LTF or MPO degranulation.

DISCUSSION Eosinophils from AA subjects contained significantly greater concentrations of intracellular EDN when compared with cells from NL or AR subjects. More EDN was released after IL-5 or GM-CSF activation when evaluated as nanograms of protein per 106 eosinophils but not when evaluated as the percentage of total cellular EDN. Therefore, although similar in their degree of EDN degranulation when activated, peripheral blood eosinophils from AA patients have a greater inflammatory capacity than those from AR subjects because of their higher intracellular content of this granule protein. Differences in assays and patient populations make it difficult to directly compare our data with those from previous reports, but our cellular EDN levels are within the range of reported values.10,12 Unlike previous reports that compared serum eosinophil granule protein levels, we focused on the intracellular concentration of a single eosinophil granule protein and its relative presence in a broad spectrum of subjects and conditions. Many studies have evaluated the relationship of serum values of eosinophil granule proteins (usually ECP) with eosinophil cell counts and disease activity, with variable results.5,6 In a study of children with asthma versus healthy children, Krug et al7 found decreased cellular ECP and eosinophil peroxidase in conjunction with increased serum granule protein levels. They suggested that these changes were caused by in vivo degranulation of activated circulating eosinophils. In contrast, Matsumoto et al8 found that a subpopulation of patients with asthma actually had low titers of serum ECP during exacerbation, making ECP a poor marker of disease activity. Finally, Pronk-Admiraal and Bartels9 reported that the process of clotting results in

FIG 1. Total cellular EDN. The bars (left to right) represent eosinophils (EOS) isolated from NL, AR, AA, AA+ICS, BAL after SBP, and blood after SBP subject groups and an HES subject. Statistics were determined by ANOVA and unpaired t test. *P < .05 versus the NL, AR, and AA+ICS groups.

low-level degranulation of ECP from eosinophils, which, in turn, correlates with peripheral blood eosinophil counts. In our subjects, plasma EDN concentrations increased in relationship to increasing peripheral eosinophil counts (NL < AR < AA < AA+ICS); these changes were not statistically significant, suggesting that EDN released in vivo by circulating eosinophils is minimal, even in AA subjects. Our low levels of circulating EDN may be due to our measurement of EDN rather than ECP or to our analyzing plasma rather than serum. We did not find any correlations among intracellular EDN concentrations, peripheral blood eosinophil counts, and plasma EDN concentrations. EDN was selected as a representative eosinophil granule protein because it has the highest soluble recovery from lysed cells (H. Kita, unpublished data). Because

572 Sedgwick et al

J ALLERGY CLIN IMMUNOL SEPTEMBER 2004

Mechanisms of asthma and allergic inflammation FIG 2. Eosinophil (EOS) degranulation. The bars (left to right) represent eosinophils isolated from NL, AR, AA, AA+ICS, BAL after SBP, and blood after SBP subject groups. A, EDN release. *P < .05 versus AR and BAL; +P < .05 versus AR and BAL and blood post-SBP; #P < .05 versus all other groups. B, EDN release measured as a percentage total intracellular EDN. *P < .05 versus NL, AR, AA, and AA+ICS; #P < .05 versus NL and AA.

the asthma-specific increase in EDN protein may reflect a generalized cellular increase in granule proteins, intracellular ECP was also measured. No significant differences in ECP were found between any of the subject groups; however, AA subjects tended to have the lowest intracellular eosinophil ECP levels. This is consistent with other studies.7,18 Our finding of decreased intracellular ECP and increased EDN may reflect differential activation of the cells and, hence, piecemeal degranulation of circulating eosinophils. Alternatively, differences in intracellular EDN and ECP levels between groups could be due to the process of cell isolation. All of the blood samples were processed identically by Percoll density fractionation and negative anti-CD16 magnetic bead selection. BAL eosinophils, however, required only density fractionation because of the extremely low number of airway neutrophils. It has been suggested that circulating and airway eosinophils from patients with asthma are functionally primed or even partially activated in vivo19,20 and are responsible for increased levels of granule proteins in serum and BAL

fluids from subjects with asthma.6,21 This rationale would suggest that peripheral eosinophils, and, even more so, BAL eosinophils isolated from patients with asthma contain less intracellular EDN rather than the increase we observed. If this is the case, our data would actually underestimate the difference between AA (blood or BAL) EDN levels and those in the other groups. Intracellular levels of neutrophil LTF and MPO were equivalent for the different subject groups. Likewise, no differences were observed in stimulated neutrophil degranulation, suggesting that increased levels of cellular granule protein may be specific to eosinophils and asthma. The differential requirement of cytochalasin B for optimal in vitro degranulation between eosinophils and neutrophils agrees with previous reports16,17 and suggests fundamental differences in the mechanisms of neutrophil and eosinophil degranulation. The higher concentrations of EDN in eosinophils from AA subjects, along with the increased numbers of circulating eosinophils, may be the result of enhanced release of immature cells from the bone marrow. However,

the morphology of the isolated peripheral eosinophils from our various subject groups appeared similar by both light and electron microscopy (data not shown). Rosenberg et al22 reported that eosinophils cultured in vitro from cord blood are slow to produce high concentrations of EDN and other granule proteins, with < 5% of the cells expressing EDN by day 14 of culture. Even after 35 to 42 days of culture, when >95% of the cells expressed characteristic eosinophil morphology, the intracellular concentrations of EDN were lower (700-1600 ng/1 3 106 cells) than the concentrations observed in any of our groups.23 Further culture of the developing cells resulted in a profound loss of EDN content. In contrast, Al-Rabia et al24 reported expression of major basic protein and ECP by 15% to 20% of cultured peripheral blood CD34+ stem cells incubated with stem cell factor, FLT1 ligand, IL-3, and IL-5 for 3 to 7 days. Because no comparisons were given in these studies for the relative EDN content of circulating, mature eosinophils, it is difficult to compare these values with our observations. To our knowledge, there is no evidence that developing eosinophils have high concentrations of granule proteins that then decrease with maturity. Our observations of similar eosinophil morphology between subject groups (data not shown) and no evidence of in vivo degranulation from the plasma EDN levels suggest that immature circulating eosinophils were not the cause of enhanced intracellular EDN in the AA group. Alternatively, the increased intracellular EDN protein in eosinophils from AA subjects could be due to enhanced mRNA transcription. Normalized to rRNA, EDN mRNA was similar in all subject groups. We have not, however, investigated the possibility that enhanced intracellular EDN protein is regulated by altered mRNA stabilization or posttranscriptional events. Eosinophils from our single HES subject had the highest EDN mRNA/rRNA ratio but the lowest concentration of intracellular EDN. This agrees with a report that eosinophils from HES patients contain significantly less EDN and ECP compared with a group of healthy subjects11 and with our finding that intracellular EDN concentrations did not correlate with mRNA levels in any of our groups. Increased intracellular EDN was limited to untreated AA patients; subjects with asthma treated with long-term (>1 year) ICS therapy had EDN levels similar to those of NL controls and AR subjects. Because in vitro treatment with corticosteroids potentiates eosinophil apoptosis, even when IL-5 or GM-CSF is present to promote cell survival, it is possible that ICS eliminated a cytokine-primed, EDNincreased, longer-surviving subpopulation of eosinophils and resulted in normalization of the intracellular EDN concentration. Neither the kinetics of the observed ICSdependent decrease in cellular EDN content nor the mechanism of its possible reversibility upon withdrawal of ICS is known, but both are currently under evaluation. Eosinophils isolated from the peripheral circulation and airway lumen 48 hours after SBP-AG had equivalent concentrations of intracellular EDN. This was unexpected because these 2 populations of eosinophils had been exposed to quite distinct in vivo compartments. Erpenbeck

et al19 reported decreased intracellular eosinophil peroxidase in BAL eosinophils compared with their corresponding circulating cells 24 hours after SBP-AG. We have reported that the increased number of BAL eosinophils 48 hours after SBP-AG strongly correlated with increased lavage fluid concentrations of eosinophil granule proteins, including EDN.21 These observations suggest that eosinophils recruited into the airway tissue are primed or activated to release some or all of their granule proteins into the lumen. Our current data, however, suggest that eosinophils isolated from the BAL fluid have migrated through the airway tissue without releasing EDN. If this is the case, what is the source of BAL fluid granule proteins? There are several possibilities. First, eosinophils isolated from the BAL fluid represent a subpopulation of infiltrating cells that have not been, or cannot be, stimulated by airway inflammatory mediators to degranulate. Second, eosinophils isolated from the BAL fluid reflect only the cells with a density between 1.085 and 1.100 g/mL. A percentage of BAL eosinophils (10% to 50%) fractionate with the mononuclear cells (macrophages, monocytes, and lymphocytes; d < 1.085 g/mL). It may be these less dense, and more vacuolated, eosinophils25 that are functionally primed or activated during recruitment and that release the granule proteins observed in BAL fluid after allergen challenge. Third, the granule proteins in the BAL lavage fluid may have diffused from degranulated eosinophils trapped in the airway tissue. Further study will be required to determine which, if any, of these possibilities apply. Although their exact role remains to be fully defined, eosinophils are a major characteristic of airway inflammation in allergic disease and asthma. In non– corticosteroid-treated asthma, however, eosinophils are phenotypically distinct in their intracellular content of EDN. Higher intracellular concentrations of granule proteins and their degranulation, combined with the increased numbers of circulating and airway eosinophils during asthma exacerbations, may be a factor in the development of airway inflammation in asthma. We thank Dr Nizar Jarjour for the BAL samples and Anne Brooks and Elizabeth Hazel for their technical help.

REFERENCES 1. World Health Initiative. Allergic rhinitis and its impact on asthma (ARIA). MSD 2001. 2. Palma-Carlos AG, Branco-Ferreira M, Palma-Carlos ML. Allergic rhinitis and asthma: more similarities than differences. Allerg Immunol (Paris) 2001;33:237-41. 3. Vignola AM, Bousquet J. Rhinitis and asthma: a continuum of disease? Clin Exp Allergy 2001;31:674-7. 4. Coyle AJ, Uchida D, Ackerman SJ, Mitzner W, Irvin CG. Role of cationic proteins in the airway. Hyperresponsiveness due to airway inflammation. Am J Respir Crit Care Med 1994;150(5 pt 2):S63-71. 5. Durham SR, Loegering DA, Dunnette S, Gleich GJ, Kay AB. Blood eosinophils and eosinophil-derived proteins in allergic asthma. J Allergy Clin Immunol 1989;84:931-6. 6. Di Lorenzo G, Drago A, Pellitteri ME, Candore G, Colombo A, Potestio M, et al. Serum levels of soluble CD23 in patients with asthma or rhinitis monosensitive to Parietaria. Its relation to total serum IgE levels and

Mechanisms of asthma and allergic inflammation

Sedgwick et al 573

J ALLERGY CLIN IMMUNOL VOLUME 114, NUMBER 3

574 Sedgwick et al

7.

8.

9.

Mechanisms of asthma and allergic inflammation

10.

11.

12.

13.

14.

15.

16.

eosinophil cationic protein during and out of the pollen season. Allergy Asthma Proc 1999;20:119-25. Krug N, Napp U, Enander I, Eklund E, Rieger CH, Schauer U. Intracellular expression and serum levels of eosinophil peroxidase (EPO) and eosinophil cationic protein in asthmatic children. Clin Exp Allergy 1999;29:1507-15. Matsumoto H, Niimi A, Minakuchi M, Izumi T. Serum eosinophil cationic protein levels measured during exacerbation of asthma: characteristics of patients with low titres. Clin Exp Allergy 2001;31:637-43. Pronk-Admiraal CJ, Bartels PC. Total amount of ECP per eosinophil as indicator for the activity state of eosinophils. Scand J Clin Lab Invest 2001;61:453-7. Horie S, Gleich GJ, Kita H. Cytokines directly induce degranulation and superoxide production from human eosinophils. J Allergy Clin Immunol 1996;98:371-81. ¨ berg G, Peterson C, Venge P. Releasability of human Carlson M, O hypereosinophilic eosinophils is related to the density of the cells. Br J Haematol 1994;86:41-7. Pronk-Admiraal CJ, Bartels PC. Eosinophil protein X concentration is dependent on eosinophil concentration. Eur J Clin Chem Clin Biochem 1997;35:351-4. Yamamoto H, Sedgwick JB, Busse WW. Differential regulation of eosinophil adhesion and transmigration by pulmonary microvascular endothelial cells. J Immunol 1998;161:971-7. Liu LY, Sedgwick JB, Bates ME, Vrtis RF, Gern JE, Kita H, et al. Decreased expression of membrane IL-5 receptor alpha on human eosinophils: I. Loss of membrane IL-5 receptor alpha on airway eosinophils and increased soluble IL-5 receptor alpha in the airway after allergen challenge. J Immunol 2002;169:6452-8. Abu-Ghazaleh RI, Dunnette SL, Loegering DA, Checkel JL, Kita H, Thomas LL, et al. Eosinophil granule proteins in peripheral blood granulocytes. J Leukoc Biol 1992;52:611-8. Huizinga TW, Dolman KM, van der Linden NJ, Kleijer M, Nuijens JH, dem Borne AE, et al. Phosphatidylinositol-linked FcRIII mediates

J ALLERGY CLIN IMMUNOL SEPTEMBER 2004

17.

18.

19.

20.

21.

22.

23.

24.

25.

exocytosis of neutrophil granule proteins, but does not mediate initiation of the respiratory burst. J Immunol 1990;144:1432-7. Hanlon WA, Stolk J, Davies P, Humes JL, Mumford R. Bonney RJ. rTNF alpha facilitates human polymorphonuclear leukocyte adherence to fibrinogen matrices with mobilization of specific and tertiary but not azurophilic granule markers. J Leukoc Biol 1991;50:43-8. Tomassini M, Magrini L, De Petrillo G, Adriani E, Bonini S, Balsano F, et al. Serum levels of eosinophil cationic protein in allergic diseases and natural allergen exposure. J Allergy Clin Immunol 1996;97:1350-5. Erpenbeck VJ, Hohlfeld JM, Petschallies J, Eklund E, Peterson CGB, Fabel H, et al. Local release of eosinophil peroxidase following segmental allergen provocation in asthma. Clin Exp Allergy 2003;33: 331-6. Carlson M, Ha˚kansson L, Peterson C, Stalenheim G, Venge P. Secretion of granule proteins from eosinophils and neutrophils is increased in asthma. J Allergy Clin Immunol 1991;87:27-33. Sedgwick JB, Calhoun WJ, Gleich GJ, Kita H, Abrams JS, Schwartz LB, et al. Immediate and late airway response of allergic rhinitis patients to segmental antigen challenge. Characterization of eosinophil and mast cell mediators. Am Rev Respir Dis 1991;144:1274-81. Rosenberg HF, Dyer KD, Li F. Characterization of eosinophils generated in vitro from CD34+ peripheral blood progenitor cells. Exp Hematol 1996;24:888-93. Ohashi H, Takei M, Ide Y, Ishii H, Kita H, Gleich GJ, et al. Effect of interleukin-3, interleukin 5 and hyaluronic acid on cultured eosinophils derived from human umbilical cord blood mononuclear cells. Int Arch Allergy Immunol 1999;118:44-50. Al-Rabia MW, Blaylock MG, Sexton DW, Thomson L, Walsh GM. Granule protein changes and membrane receptor phenotype in maturing human eosinophils cultured from CD34(+) progenitors. Clin Exp Allergy 2003;33:640-8. Sedgwick JB, Calhoun WJ, Vrtis RF, Bates ME, McAllister PK, Busse WW. Comparison of airway and blood eosinophil function after in vivo antigen challenge. J Immunol 1992;149:3710-8.