Published online before print March 10, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* School of Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; and
Cellix Ltd., Institute of Molecular Medicine, St. James Hospital, Dublin, Ireland
1Correspondence: School of Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom, AB24 2ZD. E-mail: g.m.walsh{at}abdn.ac.uk
ABSTRACT
Montelukast (MLK) is a cysteinyl leukotriene receptor-1 (cysLT1R) antagonist with inhibitory effects on eosinophils, key proinflammatory cells in asthma. We assessed the effect of MLK on resting and GM-CSF-stimulated eosinophil adhesion to recombinant human (rh)VCAM-1 at different flow rates using our novel microflow system. At 1 or 2 dyn cm–2, shear-stress unstimulated eosinophils tethered immediately to rhVCAM-1, "rolled" along part of the channel until they tethered, or rolled without tethering. At flow rates greater than 2 dyn cm–2, adherent eosinophils began to be displaced from rhVCAM-1. MLK (10 nM and 100 nM) gave partial (
40%) but significant (P<0.05) inhibition of unstimulated eosinophil adhesion to rhVCAM-1 at 1 or 2 dyn cm–2 shear stress. Once adhered, unstimulated eosinophils did not exhibit morphological changes, and GM-CSF-stimulated eosinophil adhesion under flow was characterized by greater cell flattening with significant (P<0.05) inhibition of adherent cell numbers by 100 nM MLK observed. This effect appeared specific for MLK, as the analog (E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dimethylamino)-3-oxopropyl]thio]methyl]thio]-propanoic acid, sodium salt, had no significant effect on eosinophil adhesion to VCAM-1. The possibility that LTC4, released from unstimulated or GM-CSF-treated eosinophils, contributed to their adhesion to VCAM-1 was excluded as the LT biosynthesis inhibitor 3-[1-(p-Chlorobenzyl)-5-(isopropyl)-3-t-butylthioindol-2-yl]-2,2-dimethylpropanoic acid had no inhibitory effect, and exogenously added LTC4 did not enhance eosinophil adhesion. In contrast, LTD4 enhanced eosinophil adhesion to VCAM-1, an effect blocked by MLK (10 and 100 nM). These findings demonstrate that MLK-mediated inhibition of unstimulated and GM-CSF-stimulated eosinophil adhesion to VCAM-1 under shear-stress conditions appears independent of cysLT1R antagonism.
Key Words: asthma • migration • VLA-4
INTRODUCTION
Asthma is a complex syndrome, characterized by a variable degree of airway obstruction, with many clinical phenotypes in adults and children [1
]. Modern understanding of asthma is that the fundamental abnormality is airway inflammation. There is much evidence implicating eosinophils as major effector cells in allergic-based disease, including asthma. Inappropriate accumulation of eosinophils and the subsequent release of their potent armory of mediators, including cytotoxic granule proteins, lipid mediators, cytokines, and chemokines, are thought to contribute significantly to the airway inflammation underlying asthma pathogenesis. Importantly, the increased numbers of eosinophils found in the blood, lungs, sputum, and bronchoalveolar lavage fluid of asthmatic patients correlate with disease severity, thereby strengthening the case for developing effective, eosinophil-depleting agents for clinical use [2
3
]. The preferential accumulation of eosinophils in the asthmatic lung suggests the existence of pathways for their selective accumulation [4
]. This involves enhanced eosinophil production and release from the bone marrow, followed by a combination of selective adhesion to the postcapillary vascular endothelium and transendothelial migration (TEM) into the surrounding tissues [5
]. Adhesion and TEM are mediated by the binding of leukocyte adhesion receptors to their ligands or counter-structures on the postcapillary endothelium with important contributions made by combinations of lipid mediators, chemokines, and cytokines. Early adhesive events involve the selectins, and these initiate a loose association between eosinophils and the endothelium, a process often termed "rolling". VCAM-1 and ICAM-1 regulate later and irreversible steps of TEM, leading to firm attachment and subsequent diapedesis of eosinophils. Selective eosinophil attachment to postcapillary venules occurs via binding of integrin 
4β1 to endothelial cell-bound VCAM-1 [6
7
]. TEM is then followed by chemotaxis that completes eosinophil attraction to inflammatory sites, namely, bronchial tissue in the case of asthma [2
].
Although inhaled corticosteroids remain the mainstay of anti-inflammatory therapy in asthma, they do not benefit all patients, and their use raises concerns over side-effects and compliance issues, particularly in children and adolescents [8 ]. There is therefore a clear need for new, effective asthma therapies [9 ]. The cysteinyl leukotriene cysLT1 receptor (cysLT1R) antagonists, including Montelukast (MLK), are the first new class of anti-asthma drugs to be introduced in the past 30 years. Overall, they are less effective than steroids, but some asthmatic subjects show a striking improvement, and a steroid-sparing effect has been demonstrated [10 11 12 ,]. The cysLT1R is present on the eosinophil cell surface [13 ] and may have a role in eosinophil adhesion, as stimulation of the receptor by LTD4 was shown to up-regulate eosinophil binding to recombinant human (rh)VCAM-1 [14 ]. Furthermore, MLK inhibits the transmigration of eosinophils across HUVEC under static conditions [15 ], and cysLT1 activation has been shown to increase proinflammatory cytokine release from eosinophils. MLK not only reduced this cytokine release but also down-regulated expression of cysLT1R [16 ]. More recently, MLK blocked plasma protein extravasation and eosinophil accumulation in small, intraparenchymal bronchi induced by OVA challenge of sensitized guinea pigs [17 ]. Although antagonism of the cysLT1R likely explains the majority of the anti-inflammatory properties of MLK, some effects may be independent of receptor blockade.
Our previous studies used a novel state of the art microflow system and real-time microscopic imaging to examine the inhibitory effects of a histamine H1-receptor antagonist levocetirizine on resting and granulocyte GM-CSF-stimulated human eosinophil adhesion to rhVCAM-1 under flow conditions [18 ]. There is general agreement that the introduction of flow conditions to in vitro models of leukocyte adhesion is crucial to the generation of data relevant to the in vivo situation [5 ]. Therefore, in the present study, we used the same microflow system to assess the effect of MLK on resting and GM-CSF-stimulated eosinophil adhesion to rhVCAM-1 under the precise flow rates achievable with this system, which mimic conditions within inflamed postcapillary venules.
MATERIALS AND METHODS
Materials
MLK was kindly donated by Merck and Co. Ltd. (Rahway, NJ, USA). Percoll was obtained from GE Healthcare (Buckinghamshire, UK). CD16 immunomagnetic beads and magnetically activated separation columns were from Miltenyi Biotec (Surrey, UK). VLA-4 mAb (HP2/1) and a isotype-matched control mAb were from Serotec Ltd. (Kidlington, Oxon, UK). The cysLT1R polyclonal rabbit antibody that recognizes the C-terminus of the receptor and isotype-matched control antibodies was from Cayman Chemicals (Nottingham, UK). Human GM-CSF was purchased from Peprotech (London, UK). rhVCAM-1 and human LTC4 or LTD4 were purchased from R&D Systems (Abingdon, UK). (E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dimethylamino)-3-oxopropyl]thio]methyl]thio]-propanoic acid, sodium salt (MK571) and 3-[1-(p-Chlorobenzyl)-5-(isopropyl)-3-t-butylthioindol-2-yl]-2,2-dimethylpropanoic acid (MK886) were purchased from Calbiochem (Nottingham, UK). HBSS, with or without Mg2+ (0.09767 g/ml) and Ca2+ (0.14 g/ml), was from PAA Laboratories (Yeovil, UK). BSA was purchased from Sigma (Poole, Dorset, UK). Fetal cloned serum was obtained from Fisher Scientific (Loughborough, UK). All other chemicals were of analytical grade.
Eosinophil isolation
Human eosinophils were obtained from blood donated by normal volunteers or individuals with mild allergic disease who were not taking any medication and who gave informed consent (eosinophilia 
0.5x106 eosinophils ml–1). Eosinophils were purified using our standard technique using dextran sedimentation and centrifugation on Percoll gradients followed by CD16-dependent, negative immunomagnetic selection as described [19
]. Using this method, eosinophils with a purity of at least 99% were obtained with greater than 98% viability, as assessed by trypan blue exclusion.
Microflow system
This study used a novel microflow system consisting of a novel syringe pump with a microfluidic biochip and flow sensor controlled by a PC using dedicated software (Cellix Ltd., Department of Physics, Trinity College, Dublin, Ireland). The system is described in full elsewhere [20
]. Briefly, the microfluidic syringe pump allows accurate flow rates to be achieved that are more reproducible and consistent than anything currently available. Importantly, flow rates are low (5 pL min–1–10 µL min–1), and those used in this study (1–2 dyn cm–2) mimic those found in the postcapillary venules during inflammatory processes. The biochip is comprised of eight channels, each of which mimics the dimensions of the postcapillary venules, and these are coated with the rh-adhesion protein of interest (see below). The microfluidic syringe pump is vital to the use of small diameter channels, as standard syringe pumps are incapable of delivering the required low flow rates.
Eosinophil adhesion under flow conditions
Each microchannel was coated for 1 h in humid conditions at ambient temperature with rhVCAM-1 or BSA (both 10 µg mL–1 in HBSS containing Ca2+ and Mg2+). All channels were then coated with BSA (as before) to occupy nonspecific binding sites. Resting or GM-CSF-treated eosinophils were preincubated at 37°C in a water bath for 10 min before incubation for a further 20 min with/without MLK (0.1 nM–100 nM), the analog MK571 or MK886 (both 100 nM). Anti-4β1, anti-cysLT1R polyclonal antibody, or appropriate isotype-matched antibody controls were added under the same conditions at a final concentration of 10 ug/ml. Eosinophils were then infused into the channels at stepwise increases in shear stress, from 0 to 5 dyn cm–2, 1 min per shear-stress level. Images at each shear-stress level were captured using the accompanying PixeLINK microscopy software. For experiments with GM-CSF (10 ng mL–1)- or LTC4/D4 (100 nM)-stimulated eosinophils, these were added to the warmed cells at the same time as MLK or other treatments and incubated at 37°C for 20 min prior to commencing the flow assay. Adhesion was evaluated by monitoring eosinophil migratory behavior in real time with images captured via the digital camera.
Effects of MLK on eosinophil viability and apoptosis
Viability of untreated eosinophils and eosinophils treated with a concentration range of MLK (10–5M–10–9M) was determined by trypan blue exclusion and averaged 98% and 97%, respectively. Furthermore, no concentration of MLK tested accelerated eosinophil apoptosis or inhibited GM-CSF-prolonged survival or necrosis, as judged by annexin V binding or propidium iodide exclusion, respectively (data not shown), quantified as described [21
].
Analysis of data
Three images per shear-stress level were captured, and adhered eosinophil numbers were recorded using Ducocell application software (Cellix Ltd., Trinity College). Data were exported into Excel for interpretation. Statistical significance was determined by Student’s unpaired t-test, and P < 0.05 was considered statistically significant. All data are presented as mean ± SEM. The n in the text refers to the number of separate experiments.
RESULTS
Effect of MLK on resting eosinophil adhesion to VCAM-1 under flow conditions
In total, 35,000 eosinophils were passed through each channel for the duration of each assay. Initially, 7500 cells were pumped into the channel and allowed to bind to rhVCAM-1 for 1 min without shear stress. The remaining eosinophils were pumped through the channel at step-wise increases in shear stress (1–5 dyn cm–2), and each shear stress level is applied for 1 min. Unstimulated eosinophils were shown to bind to rhVCAM-1 under no shear stress, although not all did (approximately 50% adhesion was observed, compared with <10% in the BSA-coated channel). Upon application of 1 dyn cm–2 or 2 dyn cm–2, shear-stress eosinophils were observed to tether immediately to rhVCAM-1, "roll" along part of the channel until tethering took place, or roll without tethering, leading to an increase in cells adherent to VCAM-1. No morphological changes were observed in unstimulated eosinophils once they adhered to rhVCAM-1. Eosinophils incubated with 4β1 mAb (but not an isotype control mAb) and those passed through the BSA-coated channel did not roll or tether at 1 or 2 dyn cm–2. At 3–5 dyn cm–2, resting levels of adherent eosinophils were similar to that found in the BSA-coated channel or cells preincubated with 4β1 mAb.
Figure 1
is a representative experiment that illustrates unstimulated eosinophil adhesion or GM-CSF-stimulated adhesion at 1 dyn cm–2 to VCAM-1, together with the inhibitory effects of preincubation with 100 µM MLK or anti-4β1 mAb. Comparable observations were made at 2 dyn cm–2. Dose-response studies demonstrated partial but significant inhibition (P<0.05) of resting eosinophil adhesion to rhVCAM-1 by MLK (10 nM and 100 nM) at 1 dyn cm–2 (Fig. 2A
) and 2 dyn cm–2 shear stress (Fig. 2B)
with a maximum inhibitory effect of 40%. Anti-4β1 but not an isotype control mAb significantly inhibited VCAM-1 adhesion at 1 and 2 dyn cm–2 shear stress (61.9±12.6% and 74.5±7.3% inhibition, respectively; n=4; P<0.05; Fig. 2A
and 2B
). Low levels of resting eosinophil adhesion to BSA were observed at 1 and 2 dyn cm–2 shear stress compared with VCAM-1 adhesion.
 View larger version (125K): [in a new window] |
Figure 1. Representative microscope images (200x original magnification) showing the effect of MLK on eosinophil adhesion to rhVCAM-1 (10 µg ml–1) at a flow rate of 1 dyn cm–2. Cells were preincubated in Ca2+/Mg2+-containing HBSS for 30 min at 37°C, including 20 min preincubation with MLK, 4β1 mAb, or isotype control, where appropriate, before being passed through the biochip channel (400 µm wide, 100 µm deep). Images demonstrate adhesion for unstimulated (upper panels) eosinophils and GM-CSF-stimulated eosinophils (lower panels). Isotype control mAb had no effect on unstimulated or GM-CSF-stimulated eosinophils (not shown).
|
 View larger version (11K): [in a new window] |
Figure 2. Dose response of the effect of MLK on nonstimulated eosinophil adhesion to rhVCAM-1 under flow conditions of (A) 1 dyn cm–2 and (B) 2 dyn cm–2. Data expressed as a percentage of nonstimulated eosinophil adhesion to rhVCAM-1 at 1 or 2 dyn–2, respectively. Each bar represents the mean ± SEM of at least four independent experiments. *, Significantly different (P<0.05) from control value at 1 or 2 dyn cm–2. Isotype control mAb had no effect on eosinophil adhesion (not shown).
|
4β1, or control mAb. Untreated, GM-CSF-stimulated eosinophils attached to rhVCAM-1 in the absence of shear stress and showed similar rolling and "tethering" characteristics after application of 1 or 2 dyn cm–2 shear stress. Figure 3
illustrates a modest increase in eosinophil adherence to rhVCAM-1 at 2 dyn cm–2 when preincubated with GM-CSF compared with unstimulated cells. MLK (100 nM) partially but significantly (P<0.05) inhibited GM-CSF-stimulated eosinophil adhesion to rhVCAM-1 at 2 dyn cm–2 shear stress. A significant reduction (P<0.05) of GM-CSF-stimulated eosinophil adhesion by 100 nM MLK was observed at 1 and 3–5 dyn cm–2 (data not shown). Increased, stimulated eosinophil adherence at 1 and 2 dyn cm–2 was observed with cells preincubated with anti-4β1 mAb, and cells passed through the BSA-coated channel compared with nonstimulated eosinophils (Fig. 3)
. Anti-4β1 mAb preincubation caused a significant decrease in adhesion to rhVCAM-1 by GM-CSF-stimulated cells compared with control at 2 dyn cm–2 (P<0.05; Fig. 3
) and at all shear-stress levels tested (data not shown).
 View larger version (17K): [in a new window] |
Figure 3. Dose response of the effect of MLK on GM-CSF-stimulated (10 ng mL–1, 20 min) eosinophil adhesion to rhVCAM-1 under flow conditions of 2 dyn cm–2. Data are expressed as a percentage of nonstimulated eosinophil adhesion to rhVCAM-1 at 2 dyn cm–2. Each bar represents the mean ± SEM of at least four independent experiments. *, Significantly different (P<0.05) from control value at 2 dyn cm–2. Isotype control mAb had no effect on eosinophil adhesion (not shown).
|
 View larger version (38K): [in a new window] |
Figure 4. Representative microscope images (400x original magnification) showing resting and GM-CSF-stimulated eosinophils adhesion to rhVCAM-1 (10 µg ml–1) at a flow rate of 2 dyn cm–2. Images demonstrate GM-CSF-induced flattening of eosinophils on rhVCAM-1 in the presence and absence of 100 nM MLK.
|
 View larger version (11K): [in a new window] |
Figure 5. Effect of MLK, the MLK analog MK571, the LT biosynthesis inhibitor MK886 on unstimulated eosinophil adhesion (A) or GM-CSF-stimulated adhesion (B) to rhVCAM-1 under flow conditions of 2 dyn cm–2. Data expressed as a percentage of nonstimulated eosinophil adhesion at 2 dyn cm–2. Each bar represents the mean ± SEM from four independent experiments. *, Significantly different (P<0.05) from control value at 2 dyn cm–2.
|
|
View this table: [in a new window] |
Table 1. Effect of LTC4/D4 on Eosinophil Adhesion
|
Eosinophil adhesion to postcapillary venules, a vital stage of TEM, is in part dependent on the 4β1/VCAM-1 pathway. Our previous work demonstrated inhibition of resting and GM-CSF-stimulated eosinophil adhesion to VCAM-1 by the histamine H1-receptor antagonist levocetirizine under flow conditions [18
]. In the present study, we have further analyzed resting and GM-CSF-stimulated eosinophil adhesion to VCAM-1 under different flow rates and show for the first time a reduction in resting and GM-CSF-stimulated eosinophil binding to VCAM-1 by the cysLT1R antagonist MLK. Inhibition of resting and GM-CSF-stimulated adhesion was observed at low, physiologically relevant concentrations of MLK. Significant inhibition of eosinophil adhesion to rhVCAM-1 also occurred at shear-stress levels similar to that found in inflamed, postcapillary venules, i.e., 1–2 dyn cm–2. These inhibitory effects at low drug concentrations and physiologically relevant flow rates are important findings, as they indicate the clinical relevance of MLK-mediated inhibition of VCAM-1 adhesion. If MLK-mediated inhibition of 4β1/VCAM-1-dependent eosinophil adhesion only occurred at higher concentrations or flow rates or only in the absence of shear stress, it would cast doubt as to the significance of our findings.
The inhibitory effect of MLK appears to be focused at the interaction between VCAM-1 and 4β1; i.e., it reduced initial binding of eosinophils to VCAM-1 and caused detachment of eosinophils shortly after adhesion. There did not appear to be any effect on eosinophils once they underwent flattening after attachment. This hypothesis is supported by other studies that provide evidence for a role by cysLT1R in eosinophil/VCAM-1 binding [14
15
22
]. In this study, we showed a gradual increase in GM-CSF-stimulated eosinophil adhesion to rhVCAM-1 with increasing sheer-stress levels (1–5 dyn cm–2). This compared with nonstimulated eosinophils that showed increased adhesion to rhVCAM-1 at 1–2 dyn cm–2 but decreased substantially at higher flow rates. It is thought that GM-CSF contributes to eosinophil accumulation in inflamed tissues by stimulating the proliferation and differentiation of eosinophils in the bone marrow, enhancement of eosinophil adhesion and TEM, and prolonging their persistence in the tissue via inhibition of apoptosis [2
]. As a cytokine expressed by inflamed endothelium [23
], GM-CSF may contribute to the preferential accumulation of eosinophils in vivo in part via enhancement of eosinophil TEM [24
]. GM-CSF and other cytokines appear to increase eosinophil adhesion to immobilized VCAM-1 by altering the functional state of eosinophil-expressed 4β1 from a low-affinity state to a high-affinity state rather than up-regulating 4β1 expression or receptor redistribution [25
26
]. Whether MLK inhibits GM-CSF-stimulated eosinophil adhesion by reducing 4β1 expression and interfering with receptor redistribution on the cell surface or by decreasing 4β1 affinity for VCAM-1 has yet to be determined.
cysLT1R are activated by the cysteinyl LTs LTC4, LTD4, and LTE4. LTC4 is synthesized within and released from endothelial cells and eosinophils into peripheral blood, where it is converted into LTD4 and LTE4 [16 27 ]. It is possible that autocrine release of LTC4 from unstimulated or GM-CSF-treated eosinophils may have contributed to the observed adhesion to VCAM-1. However, this possibility was excluded as the LT biosynthesis inhibitor MK886 had no inhibitory effect on unstimulated or GM-CSF-stimulated eosinophil adhesion to VCAM-1. Moreover, it has been previously demonstrated that MLK inhibits LT biosynthesis in activated leukocytes [28 ]. Importantly, we also observed that exogenously added LTC4 had no enhancing effect on eosinophil adhesion to VCAM-1 under flow in contrast to LTD4, The latter significantly enhanced eosinophil adhesion to VCAM-1 under flow, an effect blocked by MLK. Interestingly, LTD4 has been previously shown to enhance eosinophil adhesion to VCAM-1 [14 ] and eosinophil TEM through HUVEC [29 ], in both cases, under static conditions, findings that support our observation. We also observed significant inhibition of LTD4-dependent enhancement of eosinophil adhesion to VCAM-1 at 1 and 2 dyn by a cysLT1R polyclonal antibody. The latter recognizes the C-terminus of the receptor, an epitope located in the cytosol of the cell. This observation is most likely explained by uptake of the antibody into the eosinophil cytosol through pinocytosis, resulting in blockade of the epitope on the C-terminus that in turn inhibited propagation of the signal normally generated by binding of LTD4 to cysLT1R.
Two isoforms of cysLT1R exist on eosinophils: cysLT1 and cysLT2. Both receptors are coupled to G-proteins and are involved in intracellular calcium signaling [30
31
]. The role of cysLTs and their receptors in asthma and other allergic diseases has been well-documented, as has the ability of MLK to treat such conditions [10
32
33
]. This study shows that in part, the anti-inflammatory action of MLK involves interference with eosinophil 4β1/VCAM-1 binding. This could be an action independent of cysLT1 antagonism, as the MLK analog MK571 had no significant effect on unstimulated or GM-CSF-stimulated eosinophil adhesion to VCAM-1. However, further investigation is required to determine the extent to which the ability of MLK to inhibit eosinophil adhesion through the 4β1/VCAM-1 pathway is independent of an eosinophil cysLT1R blockade.
Although a significant reduction in rhVCAM-1 binding was observed in the presence of 100 nM MLK in resting and GM-CSF-stimulated eosinophils, the inhibitory effect was not as great as that seen in our previous study with the histamine H1-receptor antagonist levocetirizine [18 ]. In that study, levocetirizine not only inhibited initial binding of eosinophils to rhVCAM-1, but it also reduced GM-CSF-mediated cell flattening after adhesion, whereas MLK had no such effect. As other studies have demonstrated that MLK is effective in the treatment of asthma [11 32 33 34 35 ], this would suggest that the cysLT1R antagonism has other anti-asthmatic mechanisms of action related to eosinophils. In vivo studies have shown potential anti-asthmatic properties of MLK in mice, such as the reversal of allergen-induced airway remodeling [36 ] and reduction in VCAM-1 expression in airway vascular endothelium [22 ]. Furthermore, the cysLT1R antagonist Pranlukast inhibited human-derived eosinophil TEM across HUVEC in static conditions [37 ].
It is notable that eosinophil adhesion is high in BSA-coated channels after GM-CSF preincubation. This may be a result of eosinophil CD18/β2-integrin membrane receptors, which bind to ICAM-1 in vivo as part of TEM [38 ]. GM-CSF increases the affinity of these antigens for ICAM-1, leading to an increase in eosinophil adhesion [14 ]. β2-integrins have previously been shown to bind to BSA after IL-5 stimulation [39 ]; it is possible that stimulation of eosinophils with GM-CSF may have given a similar response in our system.
In conclusion, we have further analyzed resting and stimulated eosinophil adhesive behavior under flow conditions and demonstrate that physiologically relevant concentrations of MLK inhibit resting and GM-CSF-stimulated eosinophil adhesion to VCAM-1 under flow conditions in an in vitro model of the postcapillary venules. These demonstrate further anti-inflammatory effects by MLK that appear independent of the cysLT1R blockade, which may provide important clues for developing novel therapy aimed at blunting eosinophil-induced inflammation in allergic-based disease.
ACKNOWLEDGEMENTS
This study was supported by an unrestricted medical school grant from Merck and Co., Inc. (Rahway, NJ, USA; G. M. W.) and a FP6-2004-Mobility3 Marie Curie grant, "ASTHMA," contract number MTKI-CT-2005-029541 (V. W. and G. M. W.).
Received October 29, 2007; revised January 30, 2008; accepted February 6, 2008.
REFERENCES
4β1 (VLA-4) dependent J. Immunol. 146,3419-3423[Abstract]4β1 Integrin affinity changes govern cell adhesion J. Biol. Chem. 278,38174-38182
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
