Ligation of intercellular adhesion molecule 3 inhibits GM-CSF production by human eosinophils

Ligation of intercellular adhesion molecule 3 inhibits GM-CSF production by human eosinophils Julie M. Kessel, MD,a James E. Gern, MD,a Rose F. Vrtis, BS,b Julie B. Sedgwick, PhD,b and William W. Busse, MDb Madison, Wis

Mechanisms of allergy

Background: Intercellular adhesion molecule 3 (ICAM-3) has recently been identified on the surface of eosinophils. Objective: The purpose of this study was to characterize ICAM-3 expression on eosinophils in response to cytokines and to determine whether ligand binding of ICAM-3 modulates inflammatory responses of eosinophils, as it does in other leukocytes. Methods: To determine effects of ICAM-3 on eosinophil function, we isolated human eosinophils and used a monoclonal antibody directed against the epitope of ICAM-3 that binds to leukocyte-function antigen-1 to mimic binding of ICAM-3 and this natural ligand. We measured granulocyte-macrophage colony stimulating factor (GM-CSF) production by unstimulated eosinophils and eosinophils stimulated with ionomycin (1 mol/L), both in the presence and absence of this anti–ICAM3 antibody. Results: We found that 99% of eosinophils expressed ICAM-3, regardless of whether allergic symptoms were present or absent. Expression of ICAM-3 was not enhanced by proinflammatory cytokines. Expression of ICAM-3 was reduced in apoptotic cells and in cells incubated with the combination of GMCSF and tumor necrosis factor- (n = 3). Antibody binding of ICAM-3, which mimics leukocyte-function antigen-1 binding, had no effect on baseline GM-CSF production but reduced by 80% the production of GM-CSF stimulated by ionomycin (control 1969 pg/mL ± 1259 SD versus anti–ICAM-3 396 pg/mL ± 207 SD, n = 8) and reduced GM-CSF mRNA content. Conclusions: ICAM-3 is highly expressed on the surface of human eosinophils, and downregulation of GM-CSF production by anti–ICAM-3 mAb suggests that ICAM-3 ligation may inhibit eosinophil inflammatory responses and survival. (J Allergy Clin Immunol 2003;111:1024-31.) Key words: ICAM-3, eosinophils, adhesion molecules, cytokines, GM-CSF

Adhesion molecules on the surface of eosinophils regulate cell function and influence allergic inflammation.1,2 Intercellular adhesion molecule 3 (ICAM-3; CD50) is a

From the Departments of aPediatrics and bMedicine, University of Wisconsin Medical School. This study was supported by funds from the University of Wisconsin Graduate School. Received for publication January 7, 2002; revised January 5, 2003; accepted for publication January 14, 2003. Reprint requests: Dr J. Kessel, H4/469 CSC, University of Wisconsin Hospital, 600 Highland Ave, Madison, WI 53792-4108. © 2003 Mosby, Inc. All rights reserved. 0091-6749/2003 $30.00 + 0 doi:10.1067/mai.2003.1393

1024

Abbreviations used GM-CSF: Granulocyte-macrophage colony stimulating factor ICAM: Intercellular adhesion molecule LFA: Leukocyte-function antigen mAb: Monoclonal antibody PCR: Polymerase chain reaction

member of the ICAM family of adhesion molecules (ICAM-1,-2, -3, -4, and -5), a group of proteins that belongs to the IgG superfamily of molecules.3 ICAM-3 is highly expressed on peripheral blood monocytes, lymphocytes, and neutrophils and varies with cell maturation and activation.4-6 ICAM-3 has recently been detected on eosinophils, the predominant infiltrating cell in allergic inflammation.7-9 Involved in eosinophil adhesion to T cells, ICAM-3 may influence the role of eosinophils in allergic inflammation.8 Although it is known that ICAM1 (CD54), which is also expressed on eosinophils, can promote inflammatory responses in allergic inflammation,10-13 the role of ICAM-3 in the inflammatory potential of eosinophils is not established. Much of what is known about the role of ICAM-3 in leukocyte functions has been defined in lymphocytes and monocytes. ICAM-3 binds the 2-integrin, lymphocyte function antigen-1 (LFA-1; CD11a/CD18),4,14,15 and two newly identified ligands, CD11d/CD18 and dendritic cell–specific C-type lectin.16,17 ICAM-3 binding to LFA1 activates -integrins to a high-affinity state and can cause leukocyte-leukocyte adhesion,18,19 including eosinophil adhesion to lymphocytes.8 Moreover, like ICAM-1,20 ICAM-3 serves as a portal for “outside-in” signaling in leukocytes and induces both proinflammatory and anti-inflammatory responses. For example, ICAM-3 ligation by a monoclonal antibody (mAb) that mimics LFA-1 binding stabilizes T-cell receptor adhesion and induces IL-2 production.21,22 Similarly, ICAM3 ligation induces monocyte and neutrophil cytokine production, including MIP-1 , IL-8, and MCP-1.23 In contrast to these proinflammatory effects, ICAM-3 induces anti-inflammatory responses, for instance, by inhibiting neutrophil adhesion to endothelial cells.24 ICAM-3 also inhibits lymphocyte survival through induction of apoptosis and promotes adhesion of apoptotic leukocytes to phagocytic macrophages.25-27

Kessel et al 1025

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 5

METHODS Reagents and antibodies Percoll was obtained from Amersham Pharmacia (Uppsala, Switzerland). RPMI 1640 medium, PBS, HBSS, newborn calf serum, FCS, L-glutamine, and penicillin-streptomycin were obtained from Life Technologies (Grand Island, NY). Monoclonal mouse anti-human ICAM-3 antibody (clone CAL 3.10) and anti–ICAM-1 antibody (clone BBlG-l1) were IgG1 class (R&D Systems, Minneapolis, Minn). Irrelevant isotype-matched antibody (IgG1) raised against keyhole limpet hemocyanin (R & D Systems) was used as control. FITC-conjugated goat anti-mouse IgG was obtained from Biosource International (Camarillo, Calif). Recombinant human cytokines (IL-5, GM-CSF, IL-1 , IL-8, and TGF- ) were obtained from R & D Systems.

Subjects Leukocytes were isolated from the peripheral blood of 3 groups of subjects (age, 19 to 51 years): (1) adults who report no allergic symptoms or immediate hypersensitivity and have negative skin test results (normal control patients), (2) adults with allergic rhinitis with positive skin test results, and (3) adults with mild allergic asthma. Mild asthma was defined as a clinical diagnosis of asthma with an FEV1 ≥80% of predicted value. Immediate hypersensitivity was confirmed in each subject by at least one positive skin reaction (>3 mm) by the prick-puncture technique to common allergens (ragweed, house dust mite, grass pollen, cat dander, and dog dander). No subjects received oral or inhaled corticosteroids within 6 months of the study. The University of Wisconsin Committee for the protection of human subjects approved this study, and all participants gave written consent.

Isolation of eosinophils Eosinophils were isolated from heparinized blood as previously described.31 Briefly, after centrifugation (700g for 20 minutes) of blood over a 1.090-g/mL Percoll gradient, the mononuclear cell layer was removed. After hypotonic lysis of the red blood cell/granulocyte pellet, eosinophils were then separated from neutrophils by negative selection with the use of anti-CD16 magnetic beads (Miltenyi Biotec, Auburn, Calif). The resulting eosinophils were >95% pure and viable by trypan blue exclusion.

Surface expression of ICAM proteins on eosinophils Isolated eosinophils were incubated on ice for 30 minutes in the presence of 10 g/mL anti–ICAM-3 mAb, anti–ICAM-1 mAb, or isotype control IgG. Cells were washed with FACS buffer (PBS/2% BSA/0.2% azide), centrifuged for 5 minutes at 400g, and then incubated with FITC-conjugated goat anti-mouse IgG (1:40 dilution) for 30 minutes at 4°C. After incubation, cells were washed twice with FACS buffer, fixed with 1% paraformaldehyde, and analyzed by flow cytometry.

Cytokine regulation of surface expression of ICAM-3 on eosinophils 1

Isolated eosinophils were resuspended to a final concentration of 106 cells/mL in complete medium (RPMI 1640 with HEPES,

100 U/mL penicillin, 100 g/mL streptomycin, and 2 mmol/L Lglutamine and 10% heat–inactivated FCS). Cells were then incubated for 24 hours in 24-well tissue cultures plates in the presence or absence of cytokines, individually or in combination. Cytokine treatments included GM-CSF (1 to 100 ng/mL), TNF- (10 ng/mL), IL-10 (5 ng/mL), TGF- (4 10–10 mol/L), IL-8 (1.25 10–8 mol/L), IL-5 (10 to 100 ng/mL), and IL-1 (100 ng/mL). Eosinophils were then harvested and incubated on ice for 30 minutes in the presence of 10 g/mL FITC-conjugated anti–ICAM-3 mAb, anti–ICAM-1 mAb, or isotype control IgG. Cells were then washed with FACs buffer and incubated on ice for 15 minutes with phycoerythrin-conjugated annexin V. Propidium iodide (1 g/mL) was added just before analysis by flow cytometry.

Flow cytometry The mean fluorescence of 10,000 cells was measured with a flow cytometer (FACScan, Becton Dickinson, San Jose, Calif). Percentage of cells positive for ICAM-3 or ICAM-1 were defined as cells with a mean fluorescence >95% of the cells incubated with isotype control IgG. Surface expression of ICAM proteins was measured by using mean fluorescence units that were normalized by subtracting the mean fluorescence units of cells incubated with isotype control IgG. For cultured cells that were coincubated with phycoerythrinlabeled annexin V and propidium iodide, dead cells stained with propidium iodide were excluded from analysis. Apoptotic eosinophils, in contrast to live cells, bound annexin V and were phycoerythrin-positive.32 ICAM-3 expression on live versus apoptotic eosinophils was determined by gating on each population and then measuring the mean fluorescence (FITC) for ICAM-3.

Antibody ligation of ICAM-3 on eosinophils Isolated eosinophils were resuspended to a final concentration of 3 106 cells/mL in complete medium. Tissue culture tubes were precoated with 0.1% gelatin in HBSS for 1 hour at 37°C before eosinophils were added. The cells were then preincubated with anti–ICAM-3 mAb, anti–ICAM-1 mAb, or isotype control (10 g/mL, 15 minutes at 23°C). Ionomycin (0.25 to 2.0 mol/L) was added, and the cells were incubated at 37°C with 5% CO2 for an additional 6 to 24 hours. After culture, cell viability was determined by trypan blue exclusion.

Measurement of GM-CSF protein by ELISA After incubation of eosinophils with ionomycin in the presence or absence of anti–ICAM-3 mAb and control antibodies, cell-free supernates were stored at –70°C until assayed for GM-CSF levels by ELISA (sensitivity, 3 pg/mL; R & D Systems).

Measurement of GM-CSF mRNA by real-time polymerase chain reaction Total RNA was extracted from eosinophil cell pellets by using a 1-step phenol/chloroform extraction reagent (Tri Reagent, Sigma, St Louis, Mo) according to the manufacturer’s instructions. The total RNA was treated with Dnase (RQ1 Rnase-free Dnase, Promega, Madison, Wis) to degrade residual DNA. The total RNA was resuspended in 35 L of nuclease-free water (Promega). Reverse transcription was performed by incubating 16 L of the total RNA solution with 1 g of random primers (Promega) for 2 minutes at 70°C, followed by the addition of 400 U reverse transcriptase (Superscript II RT, Invitrogen, Carlsbad, Calif), 8 L of 5 reaction buffer, 4 L of [0.01 mol/L] DTT (Invitrogen), 80 U recombinant RNaisin (Promega), and 0.02 mol/L dNTPs (Promega), in a total volume of 42 L, with an incubation of 37°C for 1 hour, and then 94°C for 5 minutes. Next, 60 L of water was added to the cDNA. Polymerase chain reaction (PCR) amplification

Mechanisms of allergy

To improve our understanding of the biological relevance of ICAM-3 on eosinophils, we first examined the regulation of ICAM-3 expression on human eosinophils and then determined the role of this adhesion molecule in regulating the production of GM-CSF, a cytokine that promotes eosinophil activation and survival.28-30

1026 Kessel et al

J ALLERGY CLIN IMMUNOL MAY 2003

Mechanisms of allergy FIG 1. Relative human blood eosinophil, neutrophil, and mononuclear expression of ICAM-3 and ICAM-1. With the use of flow cytometry, relative ICAM-3 and ICAM-1 levels were analyzed on surfaces of eosinophils, neutrophils, and mononuclear cells isolated from peripheral blood of one subject. Representative histogram (n = 3) shows log mean fluorescence of 10,000 cells stained with anti–ICAM-3 (bold line), anti–ICAM-1 (thin line), or isotype-matched control IgG (dashed line).

detection was carried out in the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, Calif). The amplification detection was performed in a 25- L volume in a 96-well plate with an optical cover. The reaction mixture included a 1 PreDeveloped TaqMan Assay Reagent (PDAR) for GM-CSF and 1 TaqMan Universal PCR Master Mix, both from Applied Biosystems. Five microliters of cDNA template was added to each reaction. A standard curve consisting of serial dilutions of mRNA isolated from a GM-CSF–expressing cell line (SPB21) was included in the assay to calculate a relative quantity and check the assay for linearity. A cycle threshold was selected in the exponential phase of amplification. Results are given as relative PCR units.

Statistical evaluation Data are presented as mean ± SD. Data were evaluated by either a t test (parametric data), analysis of variance, or a Wilcoxon ranksum test (nonparametric data), using Sigma Stat. A value of P < .05 was accepted as significant.

RESULTS ICAM-3 expression on human blood eosinophils Human eosinophils had a high basal expression of ICAM-3 (Fig 1) that was similar to the levels observed on neutrophils and mononuclear cells. In contrast, freshly isolated eosinophils, unlike neutrophils or mononuclear cells, expressed little ICAM-1. No significant difference in ICAM-3 expression was detected on the basis of gender (7 men, 9 women) or disease classification (5 normal adults, 5 allergic asthmatic adults, and 6 adults with allergic rhinitis; data not shown). To determine whether ICAM-3 expression was altered by in vitro culture of eosinophils, flow cytometry was performed on eosinophils at the time of isolation and

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 5

Kessel et al 1027

after 24 hours of culture in medium. Eosinophils cultured in the absence of cytokines or growth factors undergo apoptosis, a process that can be detected by annexin V binding to phophatidylserine residues exposed on the surface membrane of the eosinophil.32 Basal ICAM-3 expression was similar on live eosinophils that were freshly isolated and on live cells after culture in medium for 24 hours (Fig 2). However, ICAM-3 expression was significantly reduced on apoptotic eosinophils as compared with live cells.

Cytokine regulation of ICAM-3 expression on eosinophils To determine whether ICAM-3 expression is modulated by cytokines, eosinophils were incubated with cytokines (GM-CSF, TNF- , IL-5, IL-8, TGF- , and IL10) that are known to modulate eosinophil adhesion molecule expression or inflammatory responses. Cytokine concentrations were selected on the basis of previous reports of eosinophil functional responses, including effects on adhesion molecule expression, cytokine production, and cell survival.10-12,33,34 Exposure of eosinophils to the combination of GM-CSF and TNFfor 24 hours had differential effects on ICAM-3 and ICAM-1 (Fig 3): ICAM-3 expression was reduced by approximately 44% (P = .019, n = 3), whereas ICAM-1 expression was greatly enhanced (P = .014, n = 3). Treatment for 24 hours with cytokines that induce a variety of eosinophil signaling pathways and responses (TNF- , GM-CSF, IL-10, TGF- , IL-1 , or IL-8) did not affect ICAM-3 expression (n = 3 to 5; data not shown).

Ligation of ICAM-3 by mAb inhibits eosinophil GM-CSF production Several experiments were conducted to determine whether ICAM-3 ligation affects eosinophil production of GM-CSF, an important autocrine factor that primes and

activates eosinophil inflammatory responses.28,29 First, eosinophils were incubated with increasing concentrations of ionomycin (0 to 2 mol/L) in the presence or absence of an anti–ICAM-3 mAb (10 g/mL) that mimics ligand binding by LFA-1 (mAb Cal 3.10).22,35 Ionomycin was studied because it activates autocrine production of GM-CSF, in part by increasing calcium influx and stabilizing GM-CSF mRNA.36,37 Ionomycin-induced GM-CSF production was inhibited by an anti–ICAM-3 mAb but not by isotype control antibody over an ionomycin concentration range of 0.5 to 2 mol/L, and kinetic experiments demonstrated that anti–ICAM-3 mAb inhibited GM-CSF production for at least 24 hours (n =2, data not shown). Eosinophil viability, as indicated by staining with trypan blue, was not affected by anti–ICAM-3 mAb treatment (data not shown). The effects of treatment with anti–ICAM-3 versus anti–ICAM-1 mAb on GM-CSF production were compared. Eosinophils were incubated (18 hours) with ionomycin (1 mol/L) along with either anti–ICAM-3 mAb (10 g/mL), anti–ICAM-1 mAb (10 g/mL), IgG isotype control antibody (10 g/mL), or medium alone. Eosinophil GM-CSF production was significantly inhibited by anti–ICAM-3 mAb (n = 8, P < .02), whereas anti–ICAM-1 mAb or isotype control had little effect (Fig 4). Anti–ICAM-3 mAb alone did not influence GMCSF production from eosinophils incubated in medium (IgG control, 8.4 pg/mL ± 10 SD; anti–ICAM-3, 10.2 pg/mL ± 10 SD; n = 5).

Ligation of ICAM-3 reduced GM-CSF mRNA levels in eosinophils incubated with ionomycin To determine if ICAM-3 ligation regulates GM-CSF at the level of mRNA, eosinophils were incubated with ionomycin (1 mol/L) along with anti–ICAM-3 mAb,

Mechanisms of allergy

FIG 2. Expression of ICAM-3 on live versus apoptotic eosinophils. Surface expression of ICAM-3 was determined on live cells (exclude propidium iodide and do not bind annexin V) and apoptotic cells (exclude propidium iodide and bind annexin V). Representative histogram shows ICAM-3 expression on live eosinophils at time of isolation (dashed line) and on live (thin line) and apoptotic (bold line) cells after 24-hour incubation. ICAM-3 expression was similar on live cells analyzed at time of isolation and after incubation in medium for 24 hours (n = 3). In contrast, apoptotic eosinophils (detected only after 24 hours of incubation) expressed less ICAM-3 compared with live cells (P = .002, n = 3).

1028 Kessel et al

J ALLERGY CLIN IMMUNOL MAY 2003

Mechanisms of allergy

A

B FIG 3. Cytokine-induced regulation of ICAM-3 expression on eosinophils. A, Representative histogram (n = 3) shows ICAM-3 and ICAM-1 levels on eosinophils incubated for 24 hours in medium (dashed line) compared with eosinophils treated with GM-CSF and TNF- (bold line). B, Incubation with combination of GMCSF and TNF- significantly decreased ICAM-3 expression and increased ICAM-1 expression compared with eosinophils incubated with medium (n = 3, *P ≤ .02).

anti–ICAM-1 mAb, IgG isotype control, or medium alone. After incubation (18 hours), GM-CSF mRNA levels were determined by real-time PCR. In the presence of ionomycin, ICAM-3 ligation by anti–ICAM-3 mAb decreased GM-CSF mRNA content (mean, 64%; n = 3) compared with cells incubated with IgG control (Fig 5). In contrast, anti–ICAM-1 or isotype control antibody had little effect. In the absence of ionomycin, anti–ICAM-3 mAb did not influence GM-CSF mRNA (n = 2, data not shown).

DISCUSSION The results of our study show that ICAM-3 modulates inflammatory responses in eosinophils. First, nearly all peripheral blood eosinophils expressed ICAM-3, with lower levels of ICAM-3 detected on apoptotic eosinophils. Second, in contrast to other eosinophil adhesion molecules, ICAM-3 expression was not influenced by the presence of individual cytokines but was reduced after exposure to the combination of GM-CSF and TNF. Finally, ICAM-3 caused “outside-in” signaling in eosinophils that was distinct from responses induced in

other leukocytes.22-24 A monoclonal antibody that mimics binding of LFA-1, a physiologic ligand for ICAM-3, inhibited the production of GM-CSF protein in response to ionomycin activation and reduced GM-CSF mRNA content. These findings show that ICAM-3 can modulate eosinophil function and, because of its influence on GMCSF production, may influence eosinophil survival. We studied several proinflammatory and anti-inflammatory cytokines as candidates that might be expected to modulate adhesion molecule expression on live eosinophils and found that only GM-CSF and TNF- in combination reduced the surface expression of ICAM-3. Our data show that exposure of eosinophils to a variety of proinflammatory (IL-5, IL-8, IL-1 ) or anti-inflammatory (TNF- , IL-10) cytokines did not affect ICAM-3 expression. Previous reports in lymphocytes and neutrophils showed that activation by anti–CD3 mAb, phorbol esters, or calcium ionophore decreased basal ICAM-3 levels.5,38 In contrast, ICAM-3 expression on eosinophils is not altered by phorbol esters or calcium ionophore, although expression decreases after exposure to dexamethasone.8,9 None of these studies examined cytokine

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 5

Kessel et al 1029

Mechanisms of allergy

FIG 4. Effect of ICAM-3 ligation on eosinophil GM-CSF production. Eosinophils were pretreated with 10 g/mL of anti–ICAM-3 mAb (clone Cal 3.10), anti–ICAM-1 mAb, or IgG control. Ionomycin (1 mol/L) was added to the culture. After 18-hour incubation, supernates were analyzed for GM-CSF protein. Incubation with anti–ICAM-3 mAb significantly inhibited GM-CSF production (n = 8, *P < .02).

FIG 5. Effect of ICAM-3 ligation on eosinophil GM-CSF mRNA levels. Cellular levels of mRNA for GM-CSF were measured by real-time PCR. *P < .05 for anti–ICAM-3 compared with IgG control or anti–ICAM-1.

regulation of ICAM-3, although cytokines are known to upregulate ICAM-1 expression on eosinophils.10-12 In contrast to our findings with ICAM-3, the combination of TNF- and GM-CSF increased ICAM-1 levels in eosinophils, consistent with reports by others.10-12 These data suggest that ICAM-3 and ICAM-1 may have distinct and possibly opposing actions in eosinophils. Our study did not address the mechanism of regulation of ICAM-3 expression in eosinophils, but studies in other leukocytes suggest that posttranscriptional processes decrease ICAM-3 expression. For example, ICAM-3 is proteolytically shed from the surface membrane of phorbol ester–activated mononuclear cells and neutrophils,5,38 producing a soluble form of ICAM-3 that

can also be detected in the serum of patients with active inflammatory diseases.38,39 Given that cytokines induce clustering of ICAM-3 with other surface molecules that are proteolytically shed (CD43, CD44, ICAM-1),40-42 we speculate that GM-CSF and TNF- , and potentially apoptosis, also induce ICAM-3 shedding. ICAM-3 on eosinophils serves as a portal for “outsidein” signaling, as it does in other leukocytes. We studied the effect of Cal 3.10, a monoclonal antibody that binds the “A” region of ICAM-3 and mimics LFA-1 binding.22,35 Unlike antibody studies that block receptor responses, binding of ICAM-3 by Cal 3.10 activates leukocytes. For example, Cal 3.10 induces a variety of inflammatory responses in leukocytes, including homo-

1030 Kessel et al

Mechanisms of allergy

typic aggregation and cytokine production.19,22 In eosinophils stimulated by ionomycin, we found that ligation of ICAM-3 by mAb CAL 3.10 inhibited GM-CSF protein production and decreased cellular levels of GMCSF mRNA. Under baseline conditions, Cal 3.10 had no effect on the production of the proinflammatory cytokine GM-CSF. However, when eosinophils were stimulated with ionomycin, coincubation with Cal 3.10 prevented the usual 100- to 200-fold increase in GM-CSF production caused by this ionophore. Thus, binding of LFA-1 to ICAM-3 during eosinophil interactions with other leukocytes may serve to inhibit GM-CSF under physiologic conditions that increase calcium influx, as ionomycin does, in eosinophils. Although ionophores are widely used to investigate calcium-dependent physiologic responses, the biological relevance of ionomycin-induced cytokine secretion by eosinophils requires further investigation. Ionomycin and other ionophores raise intracellular calcium, an event that occurs after exposure to physiologic agonists.43 In eosinophils, ionomycin stimulates GM-CSF production28 and activates signaling pathways, in part through its effect on GM-CSF mRNA half-life and AUUUA binding protein activity.37 Ionomycin is one approach used to study the regulation of eosinophil cytokine secretion, in part because physiologic agonists do not cause robust cytokine secretion. Recent data suggest that physiologic agonists cause GM-CSF–dependent eosinophil functions, including prolonged survival. This occurs in the absence of detectable GM-CSF protein by ELISA, in part through GM-CSF autocrine secretion and increased mRNA half-life.36,44,45 Our data suggest that ligation of ICAM-3 inhibits both the increase in GM-CSF protein production and mRNA levels in activated eosinophils. ICAM-3 binding may affect GM-CSF mRNA transcription or stabilization. Both integrins and ionomycin use posttranscriptional mechanisms to regulate eosinophil GM-CSF production by prolonging the half-life of GM-CSF mRNA, in part through interactions of 3´-untranslated region adenosineuridine (AUUUA) repeats with specific mRNA binding proteins.36,46 Given that “outside-in” signaling from ICAM-3 induces protein phosphorylation cascades,47 it is possible that ICAM-3 may affect cellular levels of GM-CSF mRNA through phosphorylation-dependent modulation of proteins that regulate mRNA transcription, stability, or decay. Alternative mechanisms for inhibition of GM-CSF production include decreased protein translation, interruption of processes involved in protein secretion, and inhibition of calcium ion influx. This report establishes an anti-inflammatory function for ICAM-3 in eosinophils through inhibition of GM-CSF production and suggests that ICAM-3 and ICAM-1 are regulated by distinct mechanisms. In our studies, blood eosinophils expressed a high ICAM-3/low ICAM-1 phenotype, consistent with previous reports.8,9 Our data also show reciprocal regulation of ICAM-3 and ICAM-1 in the presence of GM-CSF and TNF- . The biological relevance of diminished ICAM-3 levels and increased ICAM-

J ALLERGY CLIN IMMUNOL MAY 2003

1 levels, particularly given that both ICAM proteins bind to LFA-1,15 is intriguing. We speculate that high ICAM-3 expression on blood eosinophils maintains an anti-inflammatory phenotype, in part by suppressing GM-CSF production. Eosinophils exposed to cytokines appear to switch to a “lower” ICAM-3/ “higher” ICAM-1 phenotype, favoring the proinflammatory responses induced by ICAM-1, including antigen presentation, release of eosinophil-derived neurotoxin, and superoxide generation.10,12,48 In summary, we have identified ICAM-3 as one of a limited number of intrinsic inhibitors of eosinophil responses.33,34 On the basis of our findings, we suggest that ICAM-3 modulates eosinophil activation, and, if this occurs in the lung, it might be a factor in the regulation of airway inflammation. We thank Karelyn Koehn, Kristyn Jansen, and Heather Gerbyshak for technical help in the preparation of eosinophils. We also thank Lan Zeng for statistical analysis and acknowledge Chris Sorenson and David Carlton for critical reading of the manuscript.

REFERENCES 1. Schleimer RP, Bochner BS. The role of adhesion molecules in allergic inflammation and their suitability as targets of antiallergic therapy. Clin Exp Allergy 1998;28(Suppl 3):15-23. 2. Georas SN, Liu MC, Newman W, Beall LD, Stealey BA, Bochner BS. Altered adhesion molecule expression and endothelial cell activation accompany the recruitment of human granulocytes to the lung after segmental antigen challenge. Am J Respir Cell Mol Biol 1992;7:261-9. 3. Hayflick JS, Kilgannon P, Gallatin WM. The intercellular adhesion molecule (ICAM) family of proteins: new members and novel functions. Immunol Res 1998;17:313-27. 4. Vazeux R, Hoffman PA, Tomita JK, Dickinson ES, Jasman RL, St John T, et al. Cloning and characterization of a new intercellular adhesion molecule ICAM-R. Nature 1992;360:485-8. 5. del Pozo MA, Pulido R, Munoz C, Alvarez V, Humbria A, Campanero MR, et al. Regulation of ICAM-3 (CD50) membrane expression on human neutrophils through a proteolytic shedding mechanism. Eur J Immunol 1994;24:2586-94. 6. Hart SP, Ross JA, Ross K, Haslett C, Dransfield L. Molecular characterization of the surface of apoptotic neutrophils: implications for functional downregulation and recognition by phagocytes. Cell Death Differ 2000;7:493-503. 7. Gleich GJ. Mechanisms of eosinophil-associated inflammation. J Allergy Clin Immunol 2000;105:651-63. 8. Douglas IS, Leff AR, Sperling AI. CD4+ T cell and eosinophil adhesion is mediated by specific ICAM-3 ligation and results in eosinophil activation. J Immunol 2000;164:3385-91. 9. Juan M, Mullol J, Roca-Ferrer J, Fuentes M, Perez M, Vilardell C, et al. Regulation of ICAM-3 and other adhesion molecule expressions on eosinophils in vitro: effects of dexamethasone. Allergy 1999;54:1293-8. 10. Hansel TT, DeVries IJM, Carballido JM, Braun RK, Carballido-Perrig N, Rihs S, et al. Induction and function of eosinophil intercellular adhesion molecule –1 and HLA-DR. J Immunol 1992;149:2130-6. 11. Czech W, Krutmann J, Budnik A, Schopf E, Kapp A. Induction of intercellular adhesion molecule-1 (ICAM-1) expression in normal human eosinophils by inflammatory cytokines. J Invest Dermatol 1993;100:417-23. 12. Horie S, Okubo Y, Hossain M, Momose T, Suzuki J, Isobe M, et al. Intercellular adhesion molecule-1 on eosinophils is involved in eosinophil protein X release induced by cytokines. Immunology 1997;90:301-7. 13. Mengelers HJ, Maikoe T, Brinkman L, Hooibrink B, Lammers JW, Koenderman L. Immunophenotyping of eosinophils recovered from blood and BAL of allergic asthmatics. Am J Respir Crit Care Med 1994;149:345-51. 14. Fawcett J, Holness CLL, Needham LA, Turley H, Gatter KC, Mason DY, et al. Molecular cloning of ICAM-3, a third ligand for LFA-1, constitutively expressed on resting lymphocytes. Nature 1992;360:481-4. 15. Woska JR Jr, Morelock MM, Durham-Jeanfavre D, Caviness GO, Bor-

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

30.

31.

32.

mann BJ, Rothlein R. Molecular comparison of soluble intercellular adhesion molecule-1 (sICAM-1) and sICAM-3 binding to lymphocyte function-associated antigen-1. J Biol Chem 1998;273:4725-33. Van der Vieren M, Le Trong HL, Wood CL, Moore PF, St John T, Staunton DE, et al. A novel leukointegrin, d 2, binds preferentially to ICAM-3. Immunity 1995;3:683-90. Geijtenbeek TBH, Torensma R, van Vliet SJ, van Duijnhoven GCF, Adema GJ, van Kooyk Y, et al. Identification of DC-SIGN, a novel dendritic cell-specific ICAM-3 receptor that supports primary immune responses. Cell 2000;100:575-85. Cid MC, Esparza J, Juan M, Miralles A, Ordi J, Vilella R, et al. Signaling through CD50 (ICAM-3) stimulates T lymphocyte binding to human umbilical vein endothelial cells and extracellular matrix proteins via an increase in 1 and 2 integrin function. Eur J Immunol 1994;24:1377-82. Feldhaus MJ, Kessel JM, Zimmerman GA, McIntyre TM. Engagement of ICAM-3 activates polymorphonuclear leukocytes: aggregation without degranulation or 2 integrin recruitment. J Immunol 1998;161:6280-7. Hubbard AK, Rothlein R. Intercellular adhesion molecule-1 (ICAM-1) expression and cell signaling cascades. Free Radic Biol Med 2000;28:1379-86. Hayflick JS, Stine J, Fox R, Hoekstra D, Gallatin WM. Functional mapping of the cytoplasmic region of intercellular adhesion molecule-3 reveals important roles for serine residues. J Biol Chem 1997;272:22207-14. Bossy D, Buckley CD, Holness CL, Littler AJ, Murray N, Collins I, et al. Epitope mapping and functional properties of anti-intercellular adhesion molecule-3 (CD50) monoclonal antibodies. Eur J Immunol 1995;25:459-65. Kessel JM, Hayflick J, Weyrich AS, Hoffman PA, Gallatin M, McIntyre TM, et al. Coengagement of ICAM-3 and Fc receptors induces chemokine secretion and spreading by myeloid leukocytes. J Immunol 1998;160:5579-87. Skubitz KM, Campbell KD, Skubitz APN. CD50 monoclonal antibodies inhibit neutrophil activation. J Immunol 1997;159:820-8. Moffatt OD, Devitt A, Bell ED, Simmons DL, Gregory CD. Macrophage recognition of ICAM-3 on apoptotic leukocytes. J Immunol 1999; 162:6800-10. López-Briones S, Portales-Pérez DP, Baranda L, de la Fuente H, Rosenstein Y, Gonzalez-Amaro R. Stimulation through CD50 preferentially induces apoptosis of TCR1+ human peripheral blood lymphocytes. Cell Adhes Commun 1998;6:465-79. Stucki A, Hayflick JS, Sandmaier BM. Antibody engagement of intercellular adhesion molecule 3 triggers apoptosis of normal and leukaemic myeloid marrow cells. Br J Haematol 2000;108:157-66. Kita H, Ohnishi T, Okubo Y, Weiler D, Abrams JS, Gleich GJ. Granulocyte/macrophage colony-stimulating factor and interleukin 3 release from human peripheral blood eosinophils and neutrophils. J Exp Med 1991;174:745-8. Yousefi S, Blaser K, Simon HU. Activation of signaling pathways and prevention of apoptosis by cytokines in eosinophils. Int Arch Allergy Immunol 1997;112:9-12. Nagata M, Sedgwick JB, Kita H, Busse WW. Granulocyte macrophage colony-stimulating factor augments ICAM-1 and VCAM-1 activation of eosinophil function. Am J Respir Cell Mol Biol 1998;19:158-66. Yamamoto H, Sedgwick JB, Busse WW. Differential regulation of eosinophil adhesion and transmigration by pulmonary microvascular endothelial cells. J Immunol 1998;161:971-7. Walsh GM, Dewson G, Wardlaw AJ, Levi-Schaffer F, Moqbel R. A comparative study of different methods for the assessment of apoptosis and necrosis in human eosinophils. J Immunol Methods 1998;217:153-63.

Kessel et al 1031

33. Alam R, Forsythe P, Stafford S, Fukuda Y. Transforming growth factor abrogates the effects of hematopoetins on eosinophils and induces their apoptosis. J Exp Med 1994;179:1041-5. 34. Takanaski S, Nonaka R, Xing Z, O’Byrne P, Dolovich J, Jordana M. Interleukin 10 inhibits lipopolysaccharide-induced survival and cytokine production by human peripheral blood eosinophils. J Exp Med 1994;180:711-5. 35. Holness CL, Bates PA, Littler AJ, Buckley CD, McDowall A, Bossy D, et al. Analysis of the binding site on intercellular adhesion molecule 3 for the leukocyte integrin lymphocyte function-associated antigen 1. J Biol Chem 1995;270:877-84. 36. Esnault S, Malter JS. Primary peripheral blood eosinophils rapidly degrade transfected granulocyte-macrophage colony-stimulating factor mRNA. J Immunol 1999;163:5228-34. 37. Ruth JH, Esnault S, Jarzembowski JA, Malter JS. Calcium ionophore upregulation of AUUUA-specific binding protein activity is contemporaneous with granulocyte macrophage colony stimulating factor messenger RNA stabilization in AML14.3D10 cells. Am J Cell Mol Biol 1999;21:621-8. 38. Pino-Otin MR, Vinas O, de la Fuente MA, Juan M, Font J, Torradeflot M, et al. Existence of a soluble form of CD50 (intercellular adhesion molecule-3) produced upon human lymphocyte activation: present in normal human serum and levels are increased in the serum of systemic lupus erythematosus patients. J Immunol 1995;154:3015-24. 39. Littler AJ, Buckley CD, Wordsworth P, Collins I, Martinson J, Simmons DL. A distinct profile of six soluble adhesion molecules (ICAM-1, ICAM-3, VCAM-1, E-selectin, L-selectin and P-selectin) in rheumatoid arthritis. Br J Rheumatol 1997;36:164-9. 40. Campanero MR, Pulido R, Alonso JL, Pivel JP, Pimentel-Muinos FX, Fresno M, et al. Down-regulation by tumor necrosis factor-alpha of neutrophil cell surface expression of the sialophorin CD43 and the hyaluronate receptor CD44 through a proteolytic mechanism. Eur J Immunol 1991;21:3045-8. 41. Champagne B, Tremblay P, Cantin A, St Pierre Y. Proteolytic cleavage of ICAM-1 by human neutrophil elastase. J Immunol 1998;161:6398-405. 42. del Pozo MA, Nieto M, Serrador JM, Vicente-Manzanares M, Martinez C, Sanchez-Madrid F, et al. The two poles of the lymphocyte: specialized cell compartments for migration and recruitment. Cell Adhes Commun 1998;6:125-33. 43. Mandeville JTH, Maxfield FR. Calcium and signal transduction in granulocytes. Curr Opin Hematol 1996;3:63-70. 44. Esnault S, Malter JS. Minute quantities of granulocyte-macrophage colony-stimulating factor prolong eosinophil survival. J Interferon Cytokine Res 2001;21:117-24. 45. Hoontrakoon R, Chu HW, Gardai SJ, Wenzel SE, McDonald P, Fadok VA, et al. Interleukin-15 inhibits spontaneous apoptosis in human eosinophils via autocrine production of granulocyte macrophage-colony stimulating factor and nuclear factor KB activation. Am J Respir Cell Mol Biol 2002;26:404-12. 46. Esnault S, Malter JS. Granulocyte macrophage-colony-stimulating factor mRNA is stabilized in airway eosinophils and peripheral blood eosinophils activated by TNFplus fibronectin. J Immunol 2001;166:4658-63. 47. Skubitz KM, Ahmed K, Campbell KD, Skubitz APN. CD50 (ICAM-3) is phosphorylated on tyrosine and is associated with tyrosine kinase activity in human neutrophils. J Immunol 1995;154:2888-95. 48. Takashi S, Okubo Y, Horie S. Contribution of CD54 to human eosinophil and neutrophil superoxide production. J Appl Physiol 2001;91:613-22.

Mechanisms of allergy

J ALLERGY CLIN IMMUNOL VOLUME 111, NUMBER 5