CysLT1 receptor-induced human airway smooth muscle cells proliferation requires ROS generation, EGF receptor transactivation and ERK1/2 phosphorylation

Background Cysteine-containing leukotrienes (cysteinyl-LTs) are pivotal inflammatory mediators that play important roles in the pathophysiology of asthma, allergic rhinitis, and other inflammatory conditions. In particular, cysteinyl-LTs exert a variety of effects with relevance to the aetiology of asthma such as smooth muscle contraction, eosinophil recruitment, increased microvascular permeability, enhanced mucus secretion and decreased mucus transport and, finally, airway smooth muscle cells (ASMC) proliferation. We used human ASMC (HASMC) to identify the signal transduction pathway(s) of the leukotriene D4 (LTD4)-induced DNA synthesis. Methods Proliferation of primary HASMC was measured by [3H]thymidine incorporation. Phosphorylation of EGF receptor (EGF-R) and ERK1/2 was assessed with a polyclonal anti-EGF-R or anti-phosphoERKl/2 monoclonal antibody. A Ras pull-down assay kit was used to evaluate Ras activation. The production of reactive oxygen species (ROS) was estimated by measuring dichlorodihydrofluorescein (DCF) oxidation. Results We demonstrate that in HASMC LTD4-stimulated thymidine incorporation and potentiation of EGF-induced mitogenic signaling mostly depends upon EGF-R transactivation through the stimulation of CysLT1-R. Accordingly, we found that LTD4 stimulation was able to trigger the increase of Ras-GTP and, in turn, to activate ERK1/2. We show here that EGF-R transactivation was sensitive to pertussis toxin (PTX) and phosphoinositide 3-kinase (PI3K) inhibitors and that it occurred independently from Src activity, despite the observation of a strong impairment of LTD4-induced DNA synthesis following Src inhibition. More interestingly, CysLT1-R stimulation increased the production of ROS and N-acetylcysteine (NAC) abolished LTD4-induced EGF-R phosphorylation and thymidine incorporation. Conclusion Collectively, our data demonstrate that in HASMC LTD4 stimulation of a Gi/o coupled CysLT1-R triggers the transactivation of the EGF-R through the intervention of PI3K and ROS. While PI3K and ROS involvement is an early event, the activation of Src occurs downstream of EGF-R activation and is followed by the classical Ras-ERK1/2 signaling pathway to control G1 progression and cell proliferation.


Background
Cysteine-containing leukotrienes (cysteinyl-LTs), i.e. LTC 4 , LTD 4 and LTE 4 , are pivotal inflammatory mediators formed through the 5-lipoxygenase pathway of arachidonic acid and contribute to the pathogenesis of asthma [1]. In particular, cysteinyl-LTs are very potent constrictors of human bronchi not only in vitro, but also in vivo, both in normal and in asthmatic individuals [2]. Recently, the focus in asthma therapy shifted from the short-term relief of acute bronchoconstriction to the long-term management of chronic inflammation [3]. Hallmark of this process is the infiltration of inflammatory cells, predominantly eosinophils, mast cells, and lymphocytes [4], but also a significant airway remodeling [5,6]. The features of airway remodeling include subepithelial fibrosis, elevated numbers and volume of mucous cells in the epithelium, increased amounts of airway smooth muscle cells (ASMC), and increased vascularization of the airway wall [7][8][9]. Another peculiar characteristic of chronic asthma is known to be the airway hyperresponsiveness (AHR) [10], whose underlying mechanism certainly involves the hypersensitivity to G protein coupled receptors (GPCRs) contractile agonists such as carbachol, histamine, acetylcholine and cysteinyl-LTs [11].
Furthermore, numerous contractile agents have been shown to induce proliferation of ASMC in culture [12], suggesting that a persistent stimulation with contractile agonists and inflammatory agents might play an important role in triggering and sustaining airway remodeling that, in turn, contribute to AHR in asthma [3,13]. In particular, it has been suggested that LTD 4 is able to augment growth factor-induced human ASMC (HASMC) proliferation through an "atypical" CysLT-R [14] or to directly induce proliferation in cytokine primed HASMC through a classical CysLT 1 -R [15]. Moreover, very recently montelukast, a potent CysLT 1 -R antagonist clinically used in the therapy of asthma [16,17], has been demonstrated to inhibit allergen-induced airway remodeling in an in vivo mouse model of asthma [18]. However, the precise molecular basis for LTD 4 -induced HASMC growth is not known, while the mechanism underlying other spasmogens seems to vary from agonist to agonist in addition to cell type and might involve many different and sometimes parallel pathways [5,19].
It is known that LTD 4 acts through two specific GPCRs, namely CysLT 1 and CysLT 2 [20,21], which appears to be mainly coupled to G q/11 and thus to intracellular Ca 2+ elevation in recombinant systems, but also to G i/o in some natural expressing systems [22,23]. This clearly reveals heterogeneity/promiscuity of coupling for this class of receptors as already demonstrated for many other GPCRs, in particular when comparing natural and recombinant systems [24].
A number of studies have advanced the concept that GPCRs are mediators of cell growth by demonstrating their potential to activate MAPKs, particularly the ERK1/2 [25]. Indeed, recently, we have demonstrated that CysLT 1 -R is able to phosphorylate ERK1/2 and activate Ras through a pertussis toxin (PTX) sensitive G protein in the human monocyte/macrophage-like U937 cells [26]. A number of different mitogenic pathways might link GPCRs to the nucleus, some of them requiring the activation of the small GTP-binding protein Ras or, for G q coupled receptors, protein kinase C (PKC) to directly target Raf-1 [25]. In some systems, the capacity of GPCRs to transduce a mitogenic response requires a growth factor receptor, such as the epidermal growth factor receptor (EGF-R) and its tyrosine kinase activity [27]. Indeed, it has already been suggested that LTD 4 is able to transactivate PDGF receptor in mesangial cells, but this activation was somehow attributed to the activation of a CysLT 2 -R [28].
Because of their role in chronic asthma, ASMC represent a clinically relevant cell type in which to examine the effects of LTD 4 -induced activation of cell growth [12]. Both MAPKs and phosphatidylinositol-3-OH-kinase (PI3K) have been shown to mediate mitogen-induced proliferation in these cells [29]. In addition, proliferation synergy by receptor tyrosine kinase (RTK) and GPCR activation by numerous inflammatory or contractile agents have been demonstrated in HASMC, although it did not seem to be correlated to EGF-R transactivation [30][31][32].
In this report we investigated the role and the mechanism(s) by which LTD 4 induces HASMC proliferation. We show that this lipid mediator is able to induce HASMC proliferation by itself activating a CysLT 1 -R and that the mitogenic effect is dependent upon EGF-R phosphorylation. We also demonstrated that LTD 4 induces reactive oxygen species (ROS) formation. Finally, we show here that, despite the possible contribution of G q/11 -mediated pathway, both transactivation and DNA synthesis are predominantly mediated by a G i/o protein and that CysLT 1 -Rinduced activation involves, downstream of EGF-R, the classical Src-Ras-ERK signaling pathway.

Cell culture
Smooth muscle cells from human bronchi were purchased from Cambrex (Walkersville, MD) or isolated in our laboratory as previously described [33]. Briefly, macroscopically normal lung fragments were obtained at thoracotomy. Third order bronchi were removed under sterile conditions, the connective tissue and the epithelium were removed and the smooth muscle cut into pieces approximately 10 mg each. The explants were grown at 37°C in a humidified atmosphere of 5% CO 2 in Medium 199, additioned with 20% (v/v) FBS, 100 U/ml penicillin and 100 µg/ml streptomycin in 25 cm 2 culture flask. The primary isolates were positively stained with an anti-smooth muscle α-actin antibody to assess the identity of the cultures. Thereafter, cells were routinely grown in monolayers in MEM supplemented with 10% FBS, 100 U/ml penicillin and 100 µg/ml streptomycin, passaged at a 1:3 ratio in 75 cm 2 culture flask and used between the 3 rd and 8 th passage (our isolates) or between the 3 rd and the 10 th passage (purchased cells).

Binding studies
Radioligand binding was performed at equilibrium on membranes of HASMC (0.25 mg/samples) prepared as previously described [36], in a final volume of 500 µl of PBS containing 2.5 mM CaCl 2 , 2.5 mM MgCl 2 , 10 mM glycine and 20 mM penicillamine. Equilibrium binding studies were performed at 25°C for 1 h with a mixed type binding protocol, obtained by combining both saturation (0.1-1 nM of [ 3 H]ICI198,615) and competition (3 nM -1 µM of ICI198,615) protocols in a single curve [37]. Unbound ligand was separated by rapid vacuum filtration using a Brandel cell harvester and radioactivity measured by liquid scintillation counting. Analysis of binding data was performed by means of the computer program LIG-AND (see Statistical analysis). The protein content was determined by Lowry quantization assay. Non specific binding ranged between 40 and 50% of the total binding. [ 3 H]Thymidine incorporation assay was performed as previously described [31]. Briefly, cells were subcultured into 24 well plates and subconfluent cells were serum-starved for 48 hours to synchronize the entire population cell cycle. Medium was then replaced with MEM containing 1% FBS and cells stimulated with LTD 4 (two pulses at 4 hrs interval) and EGF (single pulse) at the indicated concentration for 48 hours and incubated with [ 3 H]thymidine for the last 4 hours at a final concentration of 1 µCi ml -1 . Cells were then washed twice with ice-cold phosphate buffered saline to rinse loosely associated radioactive tracer. Acid-soluble radioactivity was removed by 20 minutes treatment with 5% trichloroacetic acid at 4°C followed by a two-step washing with 95% ethanol. The acidinsoluble portion was recovered by 60 minutes digestion with 2% Na 2 CO 3 in 0.1 M NaOH. The radioactivity was then measured by liquid scintillation counting.

Phosphorylation of EGF-R
Phosphorylation of EGF-R was performed as previously described [31]. Briefly, subconfluent cells in 60-mm dishes were serum-starved for 48 h, and 2 h before stimu-lation the medium was replaced by MEM containing 0.1% fetal bovine serum. The cells were then stimulated at 37°C with LTD 4 or EGF at the concentrations indicated for 5 min. AG1478 was preincubated for 1 hour before the addition of EGF. Monolayers were then placed on ice, washed twice with phosphate buffered saline, lysed (20 mM TrisHCl pH 7.5, 1 mM dithiotreitol, 2 mM EGTA, 20 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 and the protease inhibitor complex Complete™) and sonicated four times for 15 s. The samples were subsequently diluted in Laemmli buffer, resolved by SDS-PAGE (sodium dodecyl sulfate polyacrilamide gel electrophoresis, 20 µg/lane -6% gel) and transferred to a nitrocellulose membrane. Immunoblotting was performed with a polyclonal anti-EGF-R antibody at a concentration of 0.2 µg/ml for 18 h at 4°C and immune complex was detected by chemiluminescence using horseradish peroxidase-conjugated goat anti-rabbit IgG (immunoglobulin G) as a secondary antibody. The membrane was stripped with stripping buffer and re-blotted with polyclonal anti-EGF-R antibody. ERK1/2 phosphorylation ERK1/2 phosphorylation was performed as previously described [32]. Briefly, confluent cells into 35 mm dishes were serum-starved for 48 h, preincubated with the inhibitors for the indicated time and stimulations at 37°C were terminated by addition of ice-cold lysis buffer (see above). Thereafter, the lysates were sonicated four times on ice for 15 sec, the protein content was measured and compensated for prior to SDS-PAGE. Cell lysates were solubilized by boiling at 95°C for 5 min in Laemmli buffer and subjected to electrophoresis on 15% polyacrylamide gel. The separated proteins were transferred to a nitrocellulose membrane. Membranes were then blocked for 1 hour with 5% non-fat dried milk at room temperature and then incubated overnight at 4°C with an anti-ERK2 or anti-phosphoERK1/2 monoclonal antibody at the concentration of 1 ng/ml. Next, membranes were washed extensively and incubated with horseradish peroxidaseconjugated goat anti-mouse IgG as a secondary antibody for 1 hour at room temperature. After three washes, the immunoreactive proteins were visualized by chemiluminescence.

Ras activation
Ras activation assay was performed following the affinity precipitation protocol provided by the manufacturer (Ras pull down assay kit) as previously described [26]. Briefly, cells were serum-starved overnight, treated with appropriate stimuli, and then lysed as previously described (see above). Lysates (1 mg/ml of total cell proteins in each sample) were incubated with 10 µg of Raf-1 RBD for 45 minutes at 4°C and then centrifuged for 15 seconds at 14000 × g to pellet the agarose beads. After discarding the supernatant, agarose beads were washed with 1 ml PBS and then the pellets were resuspended in 2X Laemmli sample buffer containing DTT, boiled for 5 minutes, and finally centrifuged for 15 sec at 14000 × g. The supernatant was collected and cellular proteins resolved by SDS-PAGE using 11% (w/v) acrylamide and analyzed by western blotting (see above).

ROS measurement in living cells
HASMC were loaded with 10 µM DCFH-DA in saline buffer supplemented with 0.1% of bovine serum albumin for 1 hat 37°C (stained). At the end of the incubation, cells were washed in PBS and oxidative activity was assessed as follows. The production of reactive oxygen species (ROS) was measured by the intensity of DCF emission at 525 nm (excitation 503 nm -Perkin-Elmer LS 50B) in both stained and unstained cells. Results are expressed as the difference in fluorescence (in arbitrary units, AU) calculated as AU = [I t5 -I t0 ], where I represents the intensity of fluorescence at the specified time points.

Statistical analysis and experimental design
Ligand binding studies were analyzed using LIGAND computer program. Non specific binding was calculated by LIGAND as one of the unknown parameters of the model. Selection of the best fitting model and evaluation of the statistical significance of the parameter difference was based on the F-test for the extra sum of square principle. Parameter errors are always expressed in % coefficient of variation (%CV). The curve shown was computer generated. Statistical comparison of two groups was performed using an independent t test; multiple groups were analyzed using one way ANOVA followed by either Bonferroni or Dunnett post hoc test. Data are expressed as means ± S.E.M. Each experiment was performed at least three times in triplicate (were possible) on at least two different cell lines. Basal condition refers to cells unexposed to antagonists or inhibitors and, where necessary, vehicletreated.

LTD 4 -induced increase in DNA synthesis
In order to confirm whether the different HASMC utilized expressed an LTD 4 receptor, we routinely performed a RT-PCR reaction to amplify the human CysLT 1 -R specific DNA sequence. Fig. 1A (Upper Panel) shows the expected product of 948 bp. Indeed, HASMC also express the CysLT 2 -R (Fig. 1A, Lower Panel). We performed equilibrium binding studies in membranes from HASMC using [ 3 H]ICI198,615 as labeled ligand (Fig. 1B). Computer analysis of the mixed type curve revealed the presence of two classes of binding sites, as previously reported [38]. The high affinity binding site, representative of the CysLT 1 -R [20], exhibited a K d = 0.16 nM ± 68 %CV and a  In HASMC 1 µM LTD 4 was able to produce an increase in [ 3 H]thymidine incorporation (53% ± 4.1 S.E.M. increase vs. control; n = 27), and to potentiate EGF-induced proliferation (39% ± 4.7 S.E.M. increase vs. EGF alone; n = 15). Two pulses of 1 µM LTD 4 has been utilized because it is known that LTD 4 is highly unstable and rapidly metabolized to LTE 4 , a much weaker partial agonist at the CysLT 1 -R. Furthermore, 30 minutes pretreatment with 1 µM of the two CysLT 1 -R antagonists pranlukast and zafirlukast strongly prevented LTD 4 -induced HASMC [ 3 H]thymidine incorporation (86% ± 15 S.E.M. and 75% ± 12, respectively), as well as its potentiating effect of the EGF-induced DNA synthesis (Fig. 1C-D), demonstrating that LTD 4 is mostly acting through a classical CysLT 1 -R. As expected, neither pranlukast nor zafirlukast were able to influence the mitogenic effect of EGF.

LTD 4 -mediated EGF-R phosphorylation
To investigate a possible transactivation of the EGF-R by LTD 4 in HASMC, we tested the EGF-R tyrosine kinase inhibitor AG1478. Figure 2A shows that 1 hour pretreatment with 250 nM AG1478 fully inhibited DNA synthesis induced by LTD 4 and EGF alone or by their combination.
Because we were interested in investigating whether the EGF-R can function as a downstream signaling partner of cysteinyl-LTs, therefore acting as a point of convergence for heterogeneous signaling pathways, we used immunoblot analysis to assess the phosphorylation state of this receptor following the treatment of the cells with LTD 4 . As shown in Figure 2B and 2C, EGF-R tyrosine phosphorylation was concentration-and time-dependent, starting as early as 2 min and peaking at 3 min at a concentration of 1 µM. Thus, all subsequent experiments were performed stimulating cells with 1 µM LTD 4 for 3 min, which produced an average EGF-R phosphorylation of 80% ± 7.8 S.E.M., (Fig. 3A, n = 20). Furthermore (Fig. 3B), LTD 4 also induced potentiation of EGF-stimulated autophosphorylation (47% ± 4.8 S.E.M. increase vs. 0.1 ng/ml EGF, Fig.  3A, n = 4), a result in agreement with LTD 4 potentiation of EGF-stimulated DNA synthesis. Unsurprisingly, 1 hour pretreatment with 250 nM AG1478 totally inhibited LTD 4 -induced EGF-R phosphorylation (Fig. 3C). In addition, LTD 4 effect was significantly prevented by 1 µM zafirlukast and pranlukast (~70% and ~60%, respectively) (Fig. 3D), a result again in good agreement with the CysLT 1 -R antagonist inhibition of thymidine incorporation (see above). To confirm the involvement of ERK 1/2, we measured the amount of their phosphorylated form by western blotting. Stimulation of HASMC with10 nM LTD 4 for 5 minutes (time course and concentration-response curves not shown) produced a maximal ERK1/2 phosphorylation of 162% ± 16 S.E.M. (Fig. 4B n = 12). This effect was completely inhibited by 1 µM zafirlukast and pranlukast (Fig.  4D). Furthermore (Fig. 4C), LTD 4 also induced potentiation of EGF-stimulated MAPK activation (104% ± 28 increase vs. 0.01 ng/ml EGF; Fig. 4B n = 4).

LTD 4 -induced ERK1/2 phosphorylation
To test the hypothesis that MAPK activation could be dependent upon EGF-R transactivation, we also tested the effect of AG1478 on LTD 4 -induced ERK1/2 phosphorylation and found that 1 hour pretreatment with 250 nM of the EGF-R phosphorylation inhibitor totally prevented ERK1/2 activation by EGF and LTD 4 alone or by their combination (Fig. 4E).
The time frame of EGF-R activation (3 min) is fully compatible with that of MAPK activation (5 min). Despite there is disagreement in the literature on the necessity for prolonged activation of the MAPK cascade to produce a significant mitogenic effect [41,42], we recently demonstrated that a rapid and transient activation of ERK following TP receptor stimulation translocates ERK into the nucleus as early as 2 min, where it accumulates in its active form [32]. Thus, these data suggest that that stimulation of the CysLT 1 -R in HASMCs might be important for the control of transcription factors and cell cycle re-entry, especially with other mitogenic stimuli (e.g., EGF), if efficient cell proliferation is to be achieved.

Effect of PTX on LTD 4 -induced DNA synthesis, EGF-R phosphorylation and ERK1/2 activation
To identify which class of G protein is involved in LTD 4induced [ 3 H]thymidine incorporation, we pretreated HASMC with 100 ng/ml PTX for 20 hours. Preincubation of HASMC with PTX fully abolished LTD 4 -induced DNA synthesis (Fig. 5A). On the contrary, the toxin failed to affect EGF mitogenic response, while it reverted [ 3 H]thymidine uptake level to the one induced by EGF alone. Furthermore, PTX (either 100 or 300 ng/ml) also fully inhibited LTD 4 -induced EGF-R and ERK1/2 phosphorylation, demonstrating that not only EGF-R transactivation, but also ERK1/2 phosphorylation is totally dependent   upon a G i/o protein activation ( Fig. 5B and 5C, respectively).

Effect of genistein and PP1 on LTD 4 -induced DNA synthesis, EGF-R and ERK1/2 phosphorylation, and LTD 4induced Ras activation
It is known that GPCRs might trigger activation of a protein tyrosine kinase (PTK) such as Src in order to activate MAPKs. Thus, we tested the effect of genistein, a broad spectrum PTK inhibitor, and of PP1, a specific Src kinase inhibitor, on LTD 4 -stimulated [ 3 H]thymidine incorporation in HASMC. It is clear from Figure 6A and 6B that 30 minutes pretreatment with 50 µM genistein or 1 µM PP1 strongly inhibited DNA synthesis caused by LTD 4 (> 90%), or by costimulation of EGF and LTD 4 (> 70%). Furthermore, PP1 and to a lesser extent genistein were able to inhibit the effect induced by EGF alone.
It is also known that a possible pathway leading to EGF-R transactivation by G i coupled receptors requires Src kinase activity. However, Figure 6C shows that neither genistein nor PP1 significantly affected LTD 4 -stimulated EGF-R transactivation as well as EGF-induced autophosphorylation (data not shown). At variance, both compounds inhibited LTD 4 -(~33% and > 90%, respectively) or EGFstimulated (~34% and ~86%, respectively) ERK1/2 phosphorylation in HASMC in a similar way (Fig. 6D).
Finally, through a Ras pull-down assay we directly demonstrated that 10 nM LTD 4 was able to increase the amount of Ras-GTP in HASMC (157% ± 37 S.E.M.), confirming that also in these cells LTD 4 is able to activate the small G protein Ras at a concentration relevant for its pathophysiological role in HASMC proliferation (Fig. 6E).

Effect of wortmannin and LY294002 on LTD 4 -induced DNA synthesis, EGF-R and ERK1/2 phosphorylation
It has been previously suggested that different isoforms of PI3K might be involved in the mitogenic signal induced by G i -coupled receptors. Thus, we tested two different PI3K inhibitors, i.e. wortmannin and LY29004, and found that 30 minutes pretreatment blunted LTD 4induced DNA synthesis. On the contrary, both compounds were only partially able to inhibit [ 3 H]thymidine incorporation produced by EGF alone or in combination with LTD 4 (Fig. 7A-B). Surprisingly, a partial inhibition (~65% for both compounds) was observed on EGF-R transactivation (Fig. 7C). Furthermore, both inhibitors totally ablated LTD 4 -induced ERK1/2 phosphorylation without having an effect on EGF-stimulated MAPK activation (Fig. 7D).

LTD 4 -induced increase in ROS, and effect of NAC on EGF-R transactivation and MAPK activation
Production of ROS has been noted upon growth factor stimulation of arterial smooth muscle cells [43] and bovine tracheal myocytes [44]. We show here that pretreatment for 1 h with 10 mM of the antioxidant NAC, a ROS scavenger, completely inhibits LTD 4 -induced [ 3 H]thymidine incorporation (Fig. 8A). Similarly, NAC was also able to entirely inhibit EGF-induced DNA synthesis (Fig. 8A). Thus, to test whether LTD 4 treatment induces the intracellular generation of ROS, HASMC were loaded with DCFH-DA (10 µM), and stimulated with either LTD 4 or PDGF (as internal control). This effect was specifically inhibited by a 30 min pretreatment with 1 µM of the specific CysLT 1 -R antagonist montelukast (Fig. 8B). Finally, we also show that NAC completely ablated LTD 4 -induced EGF-R phosphorylation (Fig. 8C) and MAPK activation (Fig. 8D) as well as EGF-R autophosphorylation (data not shown).

Discussion
Classical view of molecular pharmacology has always ascribed activation of proliferation to RTKs, while GPCRs have been considered involved in the activation or inhibition of enzymes and ion channels to control second messenger intracellular levels and to activate second messenger-regulated serine/threonine kinases. However, pathways for activation of nuclear transcription in response to GPCR stimulation have now been clearly delineated and involve an always-increasing number of receptors [25].
It has been previously shown that in HASMC LTD 4 is able to increase DNA synthesis [15], particularly in combination with EGF [14], without investigating the mechanism(s) through which this occurs. We demonstrate here for the first time that LTD 4 proliferative effect requires EGF-R phosphorylation through the activation of a CysLT 1 -R, PI3K, and ROS production. We also demonstrate that, both EGF-R transactivation and DNA synthesis are strongly dependent upon a G i/o protein in these cells. Finally, these LTD 4 -mediated effects were found to induce, downstream of the EGF-R, the activation of the Src-Ras-ERK1/2 pathway.
Functional data obtained with thymidine incorporation confirmed that LTD 4 is able to potentiate EGF-induced mitogenesis [14]. However, at variance with this report, LTD 4 was found to increase thymidine incorporation by itself, as already suggested by others [15]. In addition, these effects appear almost completely sensitive to the receptor antagonists zafirlukast and pranlukast, indicative of a pharmacological profile of a classical CysLT 1 -R, again in agreement with previously published data [15]. Thus, our findings seems to corroborate other two reports, in which another specific CysLT 1 -R antagonist, montelukast, was found to attenuate airway remodeling in animal model of asthma [18,45]. However, as CysLT 2 -R is expressed in HASMC [35], we cannot completely rule out the hypothesis of its partial involvement but since the pharmacological profile has been confirmed also at two other end-points (i.e. EGF-R phosphorylation and MAPK activation), we speculate that the CysLT 2 -R is not, or at most marginally, involved in the LTD 4 -mediated effects described here.
Several lines of evidence suggest that recruitment of the EGF-R tyrosine kinase is an essential step for the mitogenic stimulus induced by a number of GPCRs [27] and, sometimes, their effect has been found to be additive or synergistic to RTK-mediated growth. However, it is clear that the signaling events that mediate such augmented growth vary depending on cell line or tissue and remain to be fully characterized in many natural systems, particularly in HASMC [12]. We report here that in HASMC LTD 4 -induced DNA synthesis appears to be totally dependent on EGF-R tyrosine kinase activity. Indeed, LTD 4 is able to transactivate EGF-R in a time and dose-dependent manner, and to potentiate EGF-stimulated autophosphorylation, in agreement with the additive mechanism of potentiation of EGF-induced thymidine incorporation. These data are consistent with previous observation that cysteinyl-LTs mediate part of ovalbumin-induced lung effects in mice via EGF-R transactivation [46]. Furthermore, this effect can be predominantly ascribed to a CysLT 1 -R. As mentioned before, synergy between RTKs and GPCRs has been already reported in HASMC, despite numerous inflammatory or contractile agents, including thrombin, histamine, and carbachol were not found to cause EGF-R tyrosine phosphorylation, nor did they increase EGF-stimulated autophosphorylation [30].
Interestingly, PTX pretreatment totally inhibited LTD 4induced thymidine incorporation as well as LTD 4 -induced EGF-R transactivation, demonstrating the involvement of a G i/o protein. Thus, CysLT 1 -R, classically known to be a G q/11 -coupled receptor in many systems [21], is also coupled to a G i/o protein in HASMC. These data are in good agreement with our previous observations that LTD 4 is only marginally able to induce a cytosolic Ca 2+ transient in these cells [33], and that CysLT 1 -R is simultaneously coupled to G q/11 and G i/o proteins (promiscuous coupling) in another natural expressing systems such as the human monocyte/macrophage-like U937 cells [22].
CysLT 1 -induced EGF-R transactivation in HASMC did not seem to involve Src, as demonstrated by the inability of two different Src kinase inhibitors to significantly affect LTD 4 -induced EGF-R phosphorylation. We also excluded that EGF-R transactivation might require metalloproteinase cleavage of proHB-EGF as well as PKC activity [47]. In fact, neither the HB-EGF inhibitor CRM197, nor GF109203X, a specific PKC inhibitor, affected LTD 4induced EGF-R phosphorylation (data not shown). Conversely, EGF-R transactivation seems to require the involvement of a PI3K, likely activated by the βγ complex of G i , which, in turn, is known to activate a number of downstream signaling molecules, including different PTKs [48].
However, LTD 4 -stimulated DNA synthesis was sensitive to Src inhibition, which suggests the involvement of Src kinase downstream of EGF-R phosphorylation [49], and, in addition, the activation of Ras and ERK1/2. Indeed, a specific MEK1 inhibitor fully prevented LTD 4 -induced DNA synthesis, and inhibited the potentiating effect of LTD 4 on EGF-stimulated thymidine incorporation, indicative of the involvement of the MAPK cascade. These data have been confirmed by demonstrating that LTD 4 was able to increase ERK1/2 phosphorylation and to augment EGF-induced MAPK activation, again with a pharmacological profile characteristic of a CysLT 1 -R, and that both effects were sensitive to Src inhibition. Furthermore, LTD 4 was observed to increase the amount of active Ras (Ras-GTP), a capacity that we already demonstrated in U937 cells [26].
Consistent with a predominant role of EGF-R transactivation, AG1478 and PTX strongly reduced LTD 4 -stimulated ERK1/2 phosphorylation. Furthermore, while wortmannin and LY294002 had only a negligible effect on EGFstimulated MAPK activation, they produced a significant inhibition of LTD 4 -induced ERK1/2 phosphorylation. These data, therefore, confirm that PI3K activation is an early event upstream of EGF-R phosphorylation in the signaling cascade linking CysLT 1 -R to thymidine incorporation. Usually, the mitogenic effect caused by EGF involves the activation of a PI3K, but this pathway is parallel to and independent from ERK1/2 phosphorylation. Accordingly, the MEK1 inhibitor was only partially able to inhibit EGFinduced thymidine incorporation. These observations are in agreement with the paradigm that G i -coupled receptorand Gβγ-stimulated MAPK activation is attenuated by inhibition of PI3K acting at a point upstream of Ras, and that this pathway requires a tyrosine kinase and Raf [25].
Looking for a mechanism linking CysLT 1 -R activation and EGF-R phosphorylation, intriguingly we observed that LTD 4 rapidly increased ROS formation in HASMC, an effect specifically inhibited by the antagonist montelu-kast. Indeed, different GPCRs have been suggested to increase ROS, which, in turn, inactivate PTP that negatively control RTK activity [50], and therefore, lead to RTK transactivation [51]. While it seems clear that ROS play a role in ASMC mitogenesis, the relevant downstream effectors are not precisely known [29,52]. In this respect, very recently it has been suggested that activation of PI3K may increase ROS formation [53]. In agreement with our results, LTD 4 has been proposed to increase production of superoxide anion [54], while, more recently, zafirlukast has been shown to interfere with the release of ROS during respiratory bursts of human polymorphonuclear neutrophils or eosinophils [55,56]. Accordingly, we observed that the ROS scavenger NAC inhibited LTD 4 -induced EGF-R transactivation, MAPK activation and thymidine incorporation.

Conclusion
In conclusion, our data demonstrate that LTD 4 , the most potent bronchoconstrictor yet identified, might potentiate EGF-induced mitogenesis in HASMC, activating a CysLT 1 -R coupled to a G i protein to cause EGF-R transactivation through the intervention of PI3K and ROS. This EGF-R induced activation triggers, downstream, the classical Src-Ras-ERK1/2 pathway to control G1 progression and cell proliferation. Furthermore, because interaction between GPCR and RTK might be exceedingly complex and potentially occurs at multiple levels, we may not completely rule out the hypothesis that a second cooperative G q/11 pathway might contribute to the LTD 4 -induced effects in HASMC [57]. This possibility is the subject of ongoing investigations in our laboratory.
Taken together, these data corroborate a second, and potentially more important, role for cysteinyl-LTs in modulating cell physiology, from epithelial [40,58] to mesangial [28] or smooth muscle cells [15]. Thus, these lipid mediators produced in large quantities by the inflammatory cells infiltrating hyperresponsive airway walls, an hallmark of chronic airways disorders such as asthma and chronic obstructive pulmonary disease, might significantly contribute to smooth muscle cell hyperplasia associated with these diseases either directly, or, mainly, potentiating growth factor-induced cell proliferation. Hence, long-term inhibition of airway remodeling might disclose new and yet underestimated effects for LT modifiers in the chronic therapy of asthma. Considering the effect of the ROS scavenger also on EGF-induced proliferation, here we propose that an increase in ROS might be a key component not only of the cysteinyl-LTs enhanced proliferative response, but more generally of the airway remodeling associated with chronic asthma.