Inhibition of Allergic Inflammation in the Airways Using Aerosolized Antisense to Syk Kinase¹

Stimulation of FcγR and FcεR leads to downstream signaling events, gene transcription, mediator release, and in some cases phagocytosis. In leukocytes, cross-linking of FcR results in the activation of Src and Syk protein tyrosine kinases. These kinases associate with immunoreceptor tyrosine-based activation motifs, which serve as specific recognition sequences in the intracellular domain of FcR (1, 2, 3). Syk plays an essential role in activation of immune cells and lymphocyte development. In mast cells, this molecule is involved in regulation of multiple intracellular signaling pathways, leading to release of allergic mediators. Important downstream targets of Syk include phospholipase Cγ1, the activation of which results in Ca²⁺ mobilization and eventually in NF-AT activation. Syk also induces activation of the mitogen-activated protein kinase cascade and generation of phosphatidylinositide 3-phosphate, which in turn regulate transcription factors necessary for cytokine gene expression (4). In addition to being expressed in macrophages (5, 6) and mast cells (7, 8, 9), Syk (p72^Syk) is expressed in eosinophils (10), neutrophils (11), T cells (12, 13), and B cells (13, 14). Recently, it has been shown that Syk is also expressed in nonhemopoietic cells: human fibroblasts (15), breast epithelium (16), and rat hepatocytes (17).

Matsuda et al. (18) demonstrated that treatment of peripheral blood monocytes with Syk antisense oligodeoxynucleotides (ASO)⁴ inhibits Syk expression when compared with cells treated with scrambled ASO (Scr ASO). This inhibition correlates with the suppression of FcγR-mediated phagocytosis, suggesting that Syk plays a critical role in FcγR-mediated cellular signaling and function in monocytes/macrophages.

We have recently demonstrated that aerosolized Syk ASO in vivo suppresses Syk expression, the release of NO and TNF from macrophages, and pulmonary inflammation in an infection model of airway inflammation (19). In the current studies we used this short-term gene therapy approach to treat allergic inflammation in the airways in a Brown Norway rat model of OVA-induced asthma.

Male Brown Norway rats (Harlan Sprague Dawley, Indianapolis, IN) and male Sprague Dawley rats (Charles River Breeding Laboratories, St.-Constant, Quebec, Canada) were housed in the Health Sciences Laboratory Animal Service (University of Alberta, Edmonton, Alberta, Canada) in filter-top cages. The animals were rested for a minimum of 1 wk before experimentation. They were exposed to 12-h light/dark cycles and given food and water ad libitum.

The Brown Norway rats were sensitized to OVA i.p. as described previously (20) and used on day 21 following sensitization. The Sprague Dawley rats were infected by s.c. injection of larvae of Nippostrongylus brasiliensis (Nb) and were studied 4 wk later (21) when the inflammation in the lungs had subsided, leaving rats sensitized for re-exposure experiments. This work was approved by the University of Alberta Health Sciences Animal Policy and Welfare Committee in accordance with guidelines of the Canadian Council for Animal Care.

A stem loop rat Syk ASO was designed to interact with Syk mRNA at three sites, increasing its efficiency, as described by Matsuda et al. (18). The 60-base phosphorothioate ASO (Nucleic Acid Facility, Department of Chemistry, University of Pennsylvania, Philadelphia, PA) and the Scr ASO (Core DNA Services, University of Calgary, Calgary, Canada) were described elsewhere (19).

1,2-Dioleoyl-3-trimethylammonium-propane (DOTAP)/dioleoyl-phosphatidyl-ethanol-amine (DOPE) liposomes were prepared using a method modified from Legendre and Szoka (22), described in detail in Ref. 19 . Cationic DOTAP:DOPE liposomes formed complexes with negatively charged oligodeoxynucleotides (23) by incubating a 2.5:1 ratio of the liposomes (1.25 mg) with either Syk ASO (0.5 mg) or the Scr ASO (0.5 mg) in a final volume of 18 ml. The liposome:Syk ASO ratio was optimized as described by Stenton et al. (19).

We used exactly the same protocol, previously shown to inhibit both Syk mRNA and protein expression in rat alveolar macrophages (19). Animals were placed in plastic boxes with lids for aerosolization. Nine milliliters of saline, liposome, Scr ASO/liposome, or Syk ASO/liposome complexes were administered per rat by nebulization for 45 min using a Sidestream nebulizer (model 1200A durable; Medic-Aid, Pagham, U.K.). Twenty-four hours later the procedure was repeated, followed by a third treatment at 48 h (unless otherwise stated). Immediately after the third treatment, animals were challenged with aerosolized saline or 5% OVA in saline for 5 min. Twenty-four hours after challenge, the animals were sacrificed.

The trachea of each rat was cannulated with polyethylene tubing attached to an 18-gauge needle, and 5 ml of ice-cold PBS was massaged into the lungs 12 times as described previously (24). The PBS was aspirated into ice-cold polypropylene tubes. The first 2 ml of bronchoalveolar lavage (BAL) fluid obtained was used to measure the TNF level. The cells were kept on ice until they were washed by centrifugation at 150 × g for 20 min and resuspended in PBS, yielding a cell viability of >95% as determined by trypan blue exclusion. The isolated BAL cells were counted and cell smears were prepared using a Cytospin (Thermo Shandon, Pittsburgh, PA). In several experiments cells were also used for flow cytometry. Cell differentials were determined by counting cytospins of the BAL cells following staining with HEMA 3 reagent (Biochemical Sciences, Swedesboro, NJ).

For histopathological analyses, rats were treated as above except that BAL was not performed. Instead, the lungs were inflated by instilling 5 ml of 10% buffered formalin (Fisher Scientific, Nepean, Ontario, Canada) via the tracheotomy. The trachea was tied closed and the inflated lung was carefully removed to avoid puncturing and placed in 10% formalin for 24 h. Sample block preparation, staining, and analysis were performed by the University of Alberta Hospital’s Department of Laboratory Medicine and Pathology. Briefly, the entire lung was processed, paraffin-embedded, sectioned at 4 μm, and stained with H&E.

The whole lung sections were then assigned a score of 0–4 (0, no inflammation; 1, mild inflammation; 2, moderate inflammation; 3, severe inflammation; 4, extreme inflammation), according to the inflammation present, by a pathologist who was blind to the various treatments. The mean inflammatory scores based on the presence of hemorrhage, congestion, edema (alveolar and interstitial), and inflammation (airway lumen, airway wall, alveolar, interstitial, and perivascular) were recorded.

For counting lung eosinophils, after performing BAL the left lung was removed and fixed immediately in 10% buffered formalin. All specimens were coded, processed, and analyzed in a standardized fashion by HistoBest (Edmonton, Alberta, Canada). Multispecimen tissue blocks were prepared as previously described (25). Sections (4-μm thick) were stained by the method of Luna (26) for the specific identification of eosinophils. Cell counting was performed at a ×400 magnification in alveolar sites and throughout parenchyma, but not in microgranuloma areas, trying to avoid unrepresentative areas for the whole tissue. The number of eosinophils per cubic millimeter (n) was obtained by the following formula: n = (number of eosinophils counted per 1-mm² area of tissue examined × 1000 μm)/(2 × average size of eosinophils plus section thickness (in micrometers)), as previously described (27). The average diameter of eosinophils was 10 μm (range, 8–12 μm) and the average field area examined was 0.10 mm² (i.e., corresponding to the average volume of 0.95 ± 0.55 mm³).

BAL cells were washed with ice-cold PBS and fixed with 5% formalin for 5 min at room temperature. After fixation cells were incubated overnight in PBS containing 5% skim milk and 0.1% BSA on ice (blocking solution). The cell suspension (1 × 10⁶ cells/ml) was then preincubated with rat Fc block (BD PharMingen, Mississauga, Ontario, Canada) diluted 1/200 for 15 min and then with mAbs directly conjugated to FITC or PE for 1 h on ice. The following mAbs were used: CD11a (WT.1), CD11b/c (OX-42), CD18 (WT.3), CD49d (Mrα4-1), CD54 (1A29), CD62L (HRL1), OX-41, and OX-52 Abs (all FITC conjugated); CD3 (G4.18), CD8 (OX-8), and anti-granulocyte (RP-1) (all PE conjugated) Abs; and the corresponding isotype-matched controls. All Abs except OX-41 and OX-52 were purchased from BD PharMingen; OX-41 and OX-52 were from Serotec (Toronto, Ontario, Canada). After two washings in PBS the cells were analyzed by flow cytometry using a FACSort (BD Biosciences Mountain View, CA) with CellQuest software (BD Biosciences). Individual populations of cells were gated according to their forward and side light scatter characteristics and immunostaining for cell type-specific markers (OX-41 for alveolar macrophages, rat granulocyte marker (RP-1), and CD3 for T lymphocytes). Discrimination of eosinophils from other cells in BAL was based on their high autofluorescence in combination with small size (low forward light scatter), high granularity (high side light scatter), and the lack of expression of other cell type-specific markers.

TNF levels were measured in the supernatant of the first 2 ml of BAL. The procedure was conducted according to the manufacturer’s instructions using a commercial ELISA kit (rat TNF-α; BioSource International, Camarillo, CA). The minimal detectable level of TNF by this assay was <4 pg/ml.

We first used OVA-sensitized Brown Norway rats and challenged their isolated tracheas by perfusing OVA through the lumen of the trachea. However, OVA did not induce strong, reproducible contractions of tracheal smooth muscle. Therefore, we used a different sensitization and challenge model: Sprague Dawley rats sensitized to the helminth Nb and challenged with LPS-free Ag isolated from Nb (24). One or 10 worm equivalents (WE)/ml were used to challenge tissue.

Four weeks following sensitization to Nb, Sprague Dawley rats were treated with or without aerosol of Syk ASO/liposome complexes (250 μg Syk ASO per rat) as described above on day 1 and 24 h later on day 2 (250 μg Syk ASO per rat). On day 3 (48 h after the first Syk ASO treatment) rats were sacrificed. A segment of trachea, carefully dissected and cleaned of connective tissue (12 rings of cartilage), was mounted in a Krebs filled organ bath as described by Pavlovic et al. (28). After stabilization (1 h), the tracheas were contracted by luminal perfusion of increasing doses of acetylcholine (−log M: 8 to 2.5), resulting in a dose-response curve that indicated the concentration of acetylcholine required to produce maximal contraction. Using washout phases (1 h), the tissues were allowed to recover before being luminally perfused with Nb Ag, resulting in tracheal contraction (circular muscle of trachea). Contraction was quantified using a pressure transducer and the data were analyzed on a computer. To standardize the Ag-induced contraction among the animals, Ag challenge data were calculated and plotted as a percentage of the maximal acetylcholine contraction.

Data are expressed as mean ± SEM for n separate experiments. For all studies, statistical analysis was performed using a Mann-Whitney test (p ≤ 0.05 was considered significant).

We studied the effects of in vivo Syk ASO treatment on OVA-induced pulmonary inflammation as measured by BAL cell number and differential. OVA challenge induced a significant increase in the total cell numbers in BAL compared with sham-challengedanimals (10.1 ± 2.4 × 10⁶ vs 3.3 ± 0.6 × 10⁶, p < 0.05) (Fig. 1 A). The same large differences in cell numbers between OVA-challenged and sham-challenged animals were observed in rats treated with aerosolized liposome (10 ± 2.2 × 10⁶ vs 2.8 ± 0.4 × 10⁶, p < 0.05) and Scr ASO/liposome complex (10.2 ± 1.3 × 10⁶ vs 3.5 ± 0.6 × 10⁶, p < 0.05). By contrast, no significant increase in total BAL cell number was observed 24 h post-Ag challenge in rats treated with Syk ASO/liposome complex (4.8 ± 0.6 × 10⁶ vs 3 ± 0.4 × 10⁶, p > 0.05). Thus, compared with Scr ASO treatment, Syk ASO inhibited OVA-induced increase in total BAL cell number by 71%.

The majority of BAL cells in sham-challenged groups were macrophages (range, 93.3–98.5%), although eosinophils (0.1–3%), neutrophils (0–2.5%), and lymphocytes (1.3–1.4%) were observed as well. OVA challenge induced an increase in absolute number (Fig. 1 B) of macrophages, neutrophils, eosinophils, and lymphocytes in BAL. Remarkably, following OVA challenge, the proportion of neutrophils and eosinophils in BAL from Scr ASO-treated rats reached 23.7 ± 4.8 and 20.2 ± 4.7%, respectively, compared with 0 ± 0 and 0.1 ± 0.1% in a sham-challenged group (data not shown). Syk ASO treatment significantly inhibited neutrophilia by 74.1% and eosinophilia by 85.2% compared with BAL cells from Scr ASO-treated animals (p < 0.05). Syk ASO treatment restored the percentage of macrophages in the BAL of OVA-challenged rats to 81.4 ± 5.9%, compared with 54.2 ± 7.7% in BAL of OVA-challenged Scr ASO-treated rats.

The absolute numbers of macrophages, neutrophils, eosinophils, and lymphocytes in BAL of OVA-challenged rats treated with Syk ASO were significantly lower than in Scr ASO animals challenged with the Ag (Fig. 1 B).

Three treatments of OVA-sensitized rats with Syk ASO over 48 h significantly inhibited the influx of inflammatory cells in BAL. To determine whether two treatments over 24 h were also sufficient to inhibit pulmonary inflammation, we conducted the same experiments as above, but rats were treated once daily on 2 consecutive days with Syk ASO/liposome complex or Scr ASO/liposome complex. As positive controls for the effects of Syk ASO we treated some rats with Syk ASO/liposome complex or Scr ASO/liposome complex once daily for 3 consecutive days before sham or OVA challenge.

In animals that received two treatments with Scr ASO, the total number of BAL cells increased from 2.8 ± 0.6 × 10⁶ to 8.5 ± 2.5 × 10⁶ following OVA challenge (data not shown). When rats received two treatments with Syk ASO, OVA challenge induced a similar increase in total BAL cell number (from 2.6 ± 0.9 × 10⁶ to 10.3 ± 1.3 × 10⁶), suggesting that two treatments with Syk ASO over 24 h were insufficient to suppress pulmonary inflammation. Positive controls confirmed that three treatments with Syk ASO significantly suppressed (52.5% inhibition) OVA-induced increase in total BAL cell number (4.8 ± 0.6 × 10⁶). In agreement with these results, two treatments with Syk ASO over 24 h failed to inhibit neutrophilia and eosinophilia induced by Ag challenge (data not shown).

We studied the duration of the inhibitory effect of in vivo treatment with Syk ASO on OVA-induced pulmonary inflammation. Rats were treated three times with Syk ASO over 48 h as described above, but then were challenged with OVA immediately after the third treatment (day 3 after the beginning of the treatment), 24 h later (day 4), or 48 h later (day 5). BAL was performed and BAL samples were obtained 24 h after the Ag challenge (i.e., on days 4, 5, and 6 after the beginning of the treatment). As shown on Fig. 2, significant reductions in total BAL cell numbers, as well as in numbers of macrophages, lymphocytes, neutrophils, and eosinophils as compared with the positive control, were observed on days 4 and 5 after the Syk ASO treatment but not on day 6. This suggests that the anti-inflammatory effect of Syk ASO lasted for 2 days after three treatments over 48 h, but on day 3 it was gone.

We found a 2.2-fold increase in TNF levels in BAL of OVA-sensitized rats 24 h following Ag challenge (Fig. 3). In animals treated with Syk ASO and challenged with the Ag immediately after the last treatment (day 3, with BAL on day 4), TNF levels were significantly lower as compared with rats treated with aerosolized saline. When the BAL was delayed until the days 5 and 6 following the beginning of Syk ASO treatment (Ag challenge on days 4 and 5), the treatment did not have a significant effect on Ag-induced TNF production (Fig. 3).

We studied the effects of Syk ASO on surface expression of several adhesion molecules belonging to different families: α₄ integrin (CD49d), β₂ integrin (CD11a, CD11b/c, CD18), Ig superfamily (ICAM-1 or CD54), and L-selectin (CD62L) in BAL cells obtained from OVA-sensitized rats challenged with OVA or saline. As shown in Fig. 4, OVA challenge in sensitized animals caused an increase in the percentage of BAL leukocytes expressing CD49d, CD11a, CD11b/c, CD18, and CD54 on their surface as compared with rats challenged with saline. The increase in CD49d, CD11a, and CD18 on eosinophils and neutrophils as well as in CD11b/c and CD54 on eosinophils, neutrophils, and macrophages reached statistical significance. The percentage of neutrophils expressing CD62L decreased following the Ag challenge, indicating the shedding of L-selectin upon leukocyte activation.

Aerosolized Syk ASO applied three times over 48 h inhibited Ag-induced up-regulation of the expression of α₄ and β₂ integrins on eosinophils and neutrophils, as well as of ICAM-1 on neutrophils. In macrophages, Syk ASO treatment significantly inhibited up-regulation of the expression of both CD11b/c and CD54. The treatment also reversed Ag-induced decrease in L-selectin expression on neutrophils (Fig. 4). However, when animals were treated with OVA and Scr ASO/liposome complex, or OVA and liposome alone, there was no effect on the cellular expression of the adhesion molecules in BAL compared with treatment with OVA alone (data not shown).

We assessed whether Syk ASO treatment could suppress airway inflammation in tissues by examining the histology of rat lungs (Fig. 5). Sensitization of rats to OVA in the absence of subsequent Ag challenge did not affect the inflammatory score 4 wk later (data not shown). Similarly, OVA-sensitized and saline-challenged rats did not have significant signs of pulmonary inflammation or had mild changes. Some control animals had rare microgranulomas similar to that previously described in normal Brown Norway rats (29). The mean inflammatory score significantly increased from 1 ± 0.1 (sensitized and saline-challenged rats) to 2.5 ± 0.1 when sensitized rats were challenged with OVA (Fig. 5, A–C vs D–F). In this case, numerous microgranulomas, prominent inflammatory peribronchial and perivascular infiltration, hemorrhage, congestion, and alveolar and interstitial edema were observed. A similar magnitude of increase in inflammatory score was observed in the lungs of OVA-challenged Scr ASO-treated rats (data not shown). Moreover, using the scoring system devised, we could not detect any statistical difference in pulmonary inflammation between Syk ASO-treated and saline-treated animals challenged with OVA (Fig. 5, G–I).

However, when more detailed analysis of lung eosinophils was performed, significant differences between Syk ASO-treated and saline-treated OVA-challenged rats were revealed. The mean number of eosinophils per cubic millimeter of lung tissue in Syk ASO-treated rats was 1.7-fold lower than in saline-treated ones (sham ASO control) and was similar to those in sham-challenged rats (Fig. 6).

We used OVA-sensitized Brown Norway rats and challenged their isolated trachea by perfusing OVA through the lumen. However, a range of concentrations of OVA (1–10 μg/ml) did not contract the isolated trachea. Therefore, we applied a different sensitization and challenge model. Sprague Dawley rats were sensitized to Nb and challenged by luminal perfusion with Ag isolated from Nb as described above. The maximum acetylcholine contraction of sensitized, non-Syk ASO-treated and Syk ASO-treated rats was 2228 mg per gram of tissue (n = 13) and 2109 mg per gram of tissue (n = 4), respectively. Therefore, Syk ASO had no effect on the acetylcholine-mediated contraction of the airway, a mechanism believed not to involve Syk tyrosine kinase. As shown in Fig. 7, the sensitized isolated trachea contracted consistently to Ag challenge with 1 and 10 WE/ml (n = 6–8). Trachea from unsensitized animals did not contract following Ag challenge, but contracted to acetylcholine (maximum contraction: 2040 mg per gram of tissue), indicating that Ag-mediated contraction occurred only after sensitization.

As reported previously using this model (19), rats were treated with Syk ASO/liposome complexes two times (over a 24-h period). Because the animals were sacrificed on day 3, no third Syk ASO treatment was given on day 3. On day 3, 24 h following the second treatment, tracheas were isolated as described in Materials and Methods. Syk ASO treatment suppressed the tracheal contraction in response to 1 WE/ml by 69.3% and 10 WE/ml by 56.6% compared with sensitized and challenged rats not treated with Syk ASO (Fig. 7). Thus, Syk ASO treatment suppresses tracheal contraction in response to Ag challenge.

Syk is an important signaling molecule mediating the inflammatory response following the engagement of FcγR and FcεR (1, 2, 3). Syk induces tyrosine phosphorylation of multiple intracellular proteins participating in signaling pathways of allergic inflammation, such as Ca²⁺ mobilization and mitogen-activated protein kinase cascade (2, 4). The critical role of Syk in FcR-mediated signaling suggests that inhibition of this molecule can affect allergic inflammatory responses.

Indeed, we have recently observed that aerosolized Syk ASO-liposome complexes suppress alveolar macrophage Syk mRNA and protein expression, Syk-dependent TNF and NO release from alveolar macrophages, and pulmonary inflammation following i.v. Ag challenge in Nb-infected Sprague Dawley rats (19). Treatment of target cells with ASO requires an appropriate delivery system because ASO are anionic molecules that cross cell membranes poorly. We used cationic liposomes, DOTAP in combination with a neutral carrier lipid (DOPE), known to enhance delivery of ASO to target cells (30, 31). Our observations suggest that Syk may be a target for gene therapy in asthma and pulmonary inflammation.

To test this hypothesis, we applied the Brown Norway rat OVA-induced asthma model (32, 33, 34), using the same protocol previously shown to inhibit both Syk mRNA and protein expression (19). Aerosolized Syk ASO delivered in liposome complexes significantly inhibited pulmonary inflammation induced by Ag challenge in OVA-sensitized animals. Neither Scr ASO nor liposome alone caused this effect, eliminating the possibility of nonspecific mechanisms involved in the treatment.

Despite the inhibitory effect of Syk ASO on inflammatory cell infiltration in the bronchoalveolar space, treatment did not appear to affect major histopathological features of lung inflammation caused by Ag challenge. One possible explanation is that the scoring system used to evaluate the lung inflammation was not sensitive. More detailed histological analysis revealed that Syk ASO greatly inhibited eosinophil infiltration in lung parenchyma. Based on these observations we suggest that Syk ASO may inhibit eosinophil extravasation induced by Ag challenge. As a consequence we observed lower eosinophilia in BAL of Syk ASO-treated animals. This is significant because eosinophils are a major source of inflammatory mediators causing tissue damage and airway hyperresponsiveness in allergic asthma (35).

To complement the data on the effect of Syk ASO on airway smooth muscle responsiveness, we investigated Ag-mediated contraction of an isolated trachea from Sprague Dawley rats sensitized to Nb. Syk ASO treatment significantly suppressed Ag-mediated, but not acetylcholine-mediated, tracheal contraction. Acetylcholine contracted tracheal smooth muscle in a Syk-independent manner, whereas the effect of Nb Ag involved Syk-mediated signaling events.

Interestingly, in sensitized animals we could not induce tracheal contraction using OVA, although Ag isolated from Nb had a potent effect on tracheal smooth muscle. It appears that OVA challenge induces strong inflammatory responses in the lungs but does not directly affect tracheal smooth muscle. It is possible that inflammatory cells present in the lungs following Ag challenge release mediators potentially leading to smooth muscle contraction, but some down-regulating factors are also released and interfere with this effect. We suggest that there may be some negative regulators preventing the response of smooth muscle to the Ag challenge in this system, e.g., PGs. Thus, we have earlier demonstrated that the PG synthetase inhibitor indomethacin can suppress intestinal responsiveness to Nb Ag (36).

One hallmark of allergic asthma, namely inflammatory cell infiltration into the bronchoalveolar space, was used to evaluate the effect of different regimens of Syk ASO treatment. Three treatments with Syk ASO in a 48-h period, but not two treatments in a 24-h period, were sufficient to suppress pulmonary inflammation induced by Ag challenge. When three treatments with Syk ASO were provided, the anti-inflammatory effect of ASO lasted until day 5 after the beginning of the treatment, but on day 6 it was no longer evident. ASO appears to have a prolonged effect blocking adequate transcription and translation of Syk kinase for 48 h following its application, but after this time adequate transcription and translation are likely restored.

In another readout system, namely airway smooth muscle responsiveness in Nb-infected Sprague Dawley rats, two treatments with Syk ASO during 48 h appeared to be sufficient to suppress Ag-mediated contraction of the isolated trachea. These data agree with our previous observations that two treatments with Syk ASO during 48 h inhibited Nb-induced pulmonary inflammation (19). The model of pulmonary inflammation induced by Nb infection is apparently more sensitive to the inhibitory effect of Syk ASO compared with the Brown Norway rat OVA-induced asthma model. The effect of Syk ASO on tracheal smooth muscle contraction induced by Ag challenge can mostly be mediated by its anti-inflammatory action, but it is also possible that the antisense can have a direct effect on smooth muscle.

By what mechanisms does Syk ASO inhibit inflammatory cell infiltration into the bronchoalveolar space? Inflammatory cell infiltration into the airways is a complex process involving numerous interactions between cells and among cells and extracellular matrix proteins. These interactions, mediating leukocyte trafficking into the inflammatory site, involve several adhesion molecules (37, 38, 39, 40, 41, 42).

A crucial role of cellular adhesion molecules in accumulation of leukocytes in the airways following antigenic challenge in OVA-induced asthma in Brown Norway rats has been demonstrated using Abs specifically neutralizing β₁ integrin (VLA-4) (41), β₂ integrins (LFA-1, Mac-1) (20, 40), and ICAM-1 (41). In our study we observed a marked up-regulation by Ag challenge of α₄ and β₂ integrin and ICAM-1 expression on leukocytes in BAL. This may account for the great influx of inflammatory cells, especially eosinophils and polymorphonuclear neutrophils (PMN), into bronchoalveolar space.

The magnitude of increase in adhesion molecule expression in BAL cells depended on the cell type. In contrast to eosinophils and PMN, expression of CD11a and CD18 on the macrophage population was not significantly increased upon Ag challenge, perhaps due to the heterogeneity of the macrophage population in BAL. The latter consists of monocytes recently recruited from blood vessels and resident macrophages in the parenchyma and in the airway lumen. Interestingly, the expression of adhesion molecules is higher on PMN in BAL than on peripheral blood PMN obtained from the same animal (M. Ulanova, unpublished observations). Although we did not address this question in the present study, one might suggest that numerous cell-cell and cell-extracellular matrix interactions during the process of leukocyte migration up-regulate the expression of integrins and ICAM-1 on the cell surface.

In contrast to other adhesion molecules, L-selectin is shed upon leukocyte activation (43), and this may explain the lower percentage of CD62L⁺ neutrophils found in BAL of Ag-challenged rats compared with sham-challenged animals. In the present study we noted that Syk ASO treatment modulates the expression of adhesion molecules on BAL leukocytes, suggesting that regulation of adhesion molecules is an important component of the anti-inflammatory effect of Syk ASO. In particular, Syk ASO can affect eosinophil-dependent mechanisms in allergic asthma and inflammation by down-regulation of CD11a, CD18, and CD49d, mediating migration of eosinophils from blood vessels to the bronchoalveolar space. As a result, lower eosinophil numbers in both lung parenchyma and BAL are present in Syk ASO-treated animals compared with saline-treated animals.

The role of adhesion molecules in pulmonary inflammation is not limited to their involvement in recruitment of leukocytes into the airways. Interaction of adhesion molecules with their ligands leads to activation of intracellular signaling pathways. In particular, engagement of β₁ integrins induced tyrosine phosphorylation of Syk and activation of NF-κB that ultimately led to the expression of proinflammatory cytokines, particularly TNF (44). TNF is widely involved in regulation of leukocyte adhesion both directly inducing expression of E-selectin, P-selectin, ICAM-1, VCAM-1, and β₂ integrins, and by stimulating proinflammatory cytokine and chemokine release (36). Accordingly, we investigated levels of TNF in BAL. Our observations indicate that TNF production in BAL, induced by Ag challenge, is inhibited by Syk ASO. TNF was shown to be produced shortly after Ag challenge (45), and the amount of this cytokine that we measured in BAL is likely the net result of its accumulation and loss during the time between antigenic challenge and the BAL procedure (24 h). Our previous experiments demonstrated that Syk ASO treatment inhibits TNF production in alveolar macrophages, indicating a direct effect of ASO on TNF mRNA transcription (19). Data from the present study expand these observations in an in vivo system and suggest that aerosolized delivery of Syk ASO leads to the decrease of TNF mRNA in alveolar macrophages and perhaps other cells such as mast cells, epithelial cells, and fibroblasts. The resulting decrease in TNF production may interfere with up-regulation of adhesion molecules following Ag challenge. In contrast, Syk may act in other ways, including an effect on the transcriptional regulation of adhesion molecule expression. Our data suggest that, after Syk ASO treatment, integrin signaling might be impaired. However, dependence of the integrin signaling on Syk is still an open question. Recent observations indicated that complement-mediated phagocytosis involving integrins CD11b/CD18 and CD11c/CD18 as complement receptors was intact in Syk^−/− macrophages (2).

Another unresolved question concerns target cells for Syk ASO delivered to the airways. Alveolar macrophages express both Syk mRNA and protein (19) and may be the primary target of Syk ASO. It is interesting that breast epithelial cell lines express Syk (16), and it may be that this signaling molecule is also expressed by airway epithelial cells and is a target of Syk ASO in our model.

The present data extend our previous observations suggesting that Syk ASO may be a useful anti-inflammatory agent. Syk ASO suppresses Ag-induced tracheal contraction as well as pulmonary inflammation in a model of asthma. Modulation of the expression of adhesion molecules in BAL cells and down-regulation of production of TNF by aerosolized Syk ASO may represent important mechanisms of its anti-inflammatory action. These findings may provide a basis for a new strategy in the anti-inflammatory therapy of human asthma.

We are grateful to Dr. J. Martin for help in establishing the Brown Norway rat asthma model and Dr. T. Allen for assistance in making liposomes.