Non-IgE–dependent activation of human lung– and cord blood–derived mast cells is induced by eosinophil major basic protein and modulated by the membrane form of stem cell factor

The allergic process is initiated when antigen causes the aggregation of occupied high-affinity IgE receptors (FcεRI) expressed on the mast cell surface. This event is followed by the immediate release of proinflammatory mediators that contribute to the early phase of the allergic process.1 In addition, mast cell degranulation induces the onset of a second “late phase” reaction, 4 to 8 hours later, characterized by the tissue infiltration of inflammatory cells, mainly eosinophils.2 We have hypothesized that during the late phase of the allergic reaction mast cells interact with and are activated by eosinophils.3,4In fact, we have shown that antigen-challenged rat peritoneal mast cells release histamine following incubation with the eosinophil mediator major basic protein (MBP) by a mechanism similar to that of IgE-independent stimuli induced by cationic compounds such as compound 48/80 and substance P.4 

The importance of mast cell activation induced by eosinophil mediators has been underappreciated because isolated human lung–derived and skin-derived mast cells do not release histamine on incubation with MBP.5 According to their protease content, human lung mast cells belong to the tryptase-positive subtype of mast cells (MC_T).6 In contrast to the tryptase- and chymase-positive subtype of mast cells (MC_TC) found, for example, in the skin and intestinal submucosa,6MC_T are well-known for their unresponsiveness to non-IgE–dependent activation.7 However, mast cell populations can express significant variation in numbers, phenotype, and function, in different anatomic locations and even in the same tissue under the microenvironment influence, particularly during the course of inflammatory responses.1 Therefore, we believe that analysis of a particular mast cell population in vitro may not necessarily reflect its behavior in vivo. Indeed, we have demonstrated that the coculture of rat or human lung mast cells with 3T3 fibroblasts, a system that mimics mast cell microenvironment in vivo, prolongs mast cell survival and increases mast cell responsiveness to IgE-dependent activation.8,9 In addition, mouse mast cells derived from bone marrow, the rodent counterpart of human MC_T, acquired responsiveness to IgE-independent stimuli when in coculture with 3T3 fibroblasts.10 Moreover, under these conditions, bone marrow–derived mast cells changed their phenotype toward connective tissue mast cells as exemplified by the increase in their histamine content and a marked increase in biosynthesis of proteoglycans that bear heparin.11 

In the present study, we investigated whether highly purified human lung mast cells in coculture with human lung fibroblasts become responsive to MBP. Mast cell responsiveness to eosinophil mediators is potentially novel evidence for an alternative pathway of allergen-independent activation able to contribute to disease. We also compared the susceptibility of cord blood–derived mast cells (CBMCs) to IgE-independent activation in suspension and in coculture conditions, and we defined a novel role for the fibroblast-derived membrane stem cell factor (SCF) in this effect.

Mononuclear cells were isolated from umbilical cord blood as previously described.12 Cells were seeded at 10⁶ cells/mL minimum essential medium (MEM-α) containing 10% (vol/vol) heat-inactivated fetal calf serum (FCS), 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mMl-glutamine, 10 μg/mL ribonucleases (Biological Industries, Beith Haemek, Israel), 100 ng/mL SCF (a generous gift from Amgen, Thousand Oaks, CA), 10 ng/mL interleukin 6 (IL-6; Peprotech, Rocky Hill, NJ), and 3 × 10⁻⁷ M prostaglandin E₂ (PGE₂; Sigma Chemicals, St Louis, MO). Half of the culture medium was replaced every week. CBMCs were harvested for the experiments between 6 and 9 weeks of culture when more than 95% of the cells were stained metachromatically with toluidine blue and positively stained for both tryptase and chymase as assessed by flow cytometry.

Human lung mast cells were isolated from normal tissue of patients undergoing surgery for lung emphysema and carcinoma using an enzymatic method.13 Human lung mast cells were purified up to 90% via selection of c-kit⁺ cells by magnetic cell separation (Dynalbeads, Dynal, Oslo, Norway). Human lung mast cells were cultured at a concentration of 10⁶/mL in Dulbecco modified Eagle medium (DMEM; Biological Industries) containing 10% vol/vol FCS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM l-glutamine (DMEM⁺) and supplemented with 100 ng/mL SCF.

Written consent was obtained from all the volunteers according to the guidelines established by the Tel Hashomer (cord blood) and Assaf Harofeh (lung biopsies) Medical Center Human Experimentation Helsinki Committees, Israel.

Rat peritoneal mast cells were obtained from Sabra rats by peritoneal lavage and purified as previously described (metrizamide gradient, 22.5%, > 96% purity).8 The rats were cared for according to the Guidelines of the Animal Committee of the Hebrew University of Jerusalem, Israel.

Human lung mast cells and CBMCs were seeded on confluent monolayers of the human lung fibroblast cell line (MRC 5) and Swiss albino mouse embryonic 3T3 fibroblast cell line (3T3; American Type Culture Collection, Rockville, MD), respectively, in 24 wells at a density of 5 × 10⁴/0.5 mL DMEM⁺. For some experiments CBMCs were seeded on either Transwell membranes (0.4-μm pore size; Nalge Nunc International, Naperville, IL) to separate them from 3T3 fibroblasts, or embryonic fibroblasts fromSl/Sl^d mutant mice (kindly provided by Dr T. Jippo, Department of Pathology, Osaka University Medical School, Japan), or 3T3 fibroblasts treated overnight with SCF sense and antisense. Culture media of CBMCs was always supplemented with 100 ng/mL SCF. Human mast cells were maintained in coculture for 4 days prior to activation. This coculture time was determined as optimal in preliminary kinetics experiments.

Oligodeoxynucleotides end-capped with 2′-O-methyl RNA substitutions at three 3′-terminal positions were used at 0.1 pM by preincubating 3T3 fibroblast overnight before and during CBMC coculture. This antisense concentration was shown to significantly inhibit the level of membrane SCF protein as assessed by confocal microscopy. Sense oligonucleotide at the same concentration did not affect SCF expression. Sequences for sense and antisense oligodeoxynucleotides have been previously published elsewhere.14 During antisense therapy, 100 ng/mL SCF was added to CBMCs cocultured with 3T3 fibroblasts to support CBMC viability. Therefore, in these cultures only the influence of the membrane form of SCF on CBMCs was selectively suppressed.

MBP was purified from eosinophils obtained from patients with marked eosinophilia as described.15,16 The purified protein was stored at −70°C, and samples were thawed immediately before use. MBP was judged pure by its banding pattern on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) after staining with Coomassie brilliant blue R. Protein concentrations were determined using the appropriate E277 values.

Cultures were incubated in Tyrode buffer containing 0.1% gelatin (Tg), 1.8 mM CaCl₂, 0.9 mM MgCl₂(Tg⁺⁺) for 30 minutes with one of the following: 2 μg/mL anti-FcεRI-α chain 22E7 (kindly provided by Dr J. P. Kochan, Hoffman La Roche, Nutley, NJ); 2, 5, and 10 μg/mL compound 48/80 (Sigma Chemicals, Rehovot, Israel); MBP at 0.001, 0.01, 0.1, 0.1, and 10 μM for human lung mast cells, and at 0.001, 0.01, 0.1, 1, 2, and 5 μM for CBMCs. To assess the effects of pertussis toxin (Ptx), cultures were incubated for 2 hours in Tg⁺⁺containing 1 μg/mL Ptx (Sigma, Israel) and then activated as described. Histamine was measured in the culture supernatants and in sonicated cells by a radioenzymatic assay.17 Histamine release was calculated as a percentage as follows: histamine in supernatant/(histamine in supernatant + histamine in cells) × 100. Generation of cysteinyl-leukotrienes (cys-LTs) in supernatants was measured with an enzyme-linked immunosorbent assay (ELISA) for LTC₄/D₄/E₄ (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom). For this ELISA, the cross-reactivity with LTC₄ is 100%, LTD₄is 100%, LTE₄ is 70%, and LTB₄ is 0.3%. Generation of prostaglandin D₂ (PGD₂) in the supernatants was measured with an ELISA for PGD₂ (Cayman Chemical, Ann Arbor, MI). For the PGD₂ ELISA, the cross-reactivity with thromboxane B₂(TXB₂), PGF_2a, or PGE₂ is less than 0.01%. Cys-LT and PGD₂ concentrations were calculated according to manufacturer's instructions and the results expressed in ng/10⁶ cells.

Mast cell viability was assessed by trypan blue exclusion test. Mast cells were examined blindly 10 minutes, 60 minutes, and 24 hours after incubation with the different activators immediately following addition of 0.1 mL trypan blue (0.4%; Sigma, Israel). The percentage of viable cells was calculated as follows: percentage of viable mast cells = (no. of trypan blue-negative cells/no. of total cells) × 100. Viability of mast cells was always more than 85% in the experiments.

Lysates were prepared from 3 × 10⁶ cells as described.18 After protein concentration was determined by Bradford method,19 samples were resolved in 10% SDS-PAGE under reducing conditions and transferred to polyvinylidene difluoride membranes. G_i3α purified from bacterial lysates was used as positive control (Santa Cruz Biotechnologies, Santa Cruz, CA). The membranes were incubated for 1 hour at room temperature with 1 μg/mL rabbit antihuman G_i3α (Santa Cruz Biotechnologies). After the membranes were washed, the proteins were detected with secondary immunopure goat antirabbit antibodies conjugated with horseradish peroxidase (1:5000; Pierce, Rockford, IL) followed by enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech). Bands were scanned and their density calculated as follows: band area × (band intensity of the sample − background intensity).

Fibroblasts were seeded on 12-mm cover glasses until confluency. Human lung mast cells and CBMCs (10⁶ cells/0.5 mL) were seeded on these fibroblasts and maintained in coculture as described. Control cells consisted of rat peritoneal mast cells that constitutively express G_i3α. Cells were fixed and permeabilized as previously described.19 Staining was performed by incubating the cells first with 5 μg/mL rabbit antihuman G_i3α (Santa Cruz Biotechnologies) for 1 hour at room temperature. After washings with phosphate-buffered saline (PBS), slides were incubated with 20 μg/mL fluorescein isothiocyanate (FITC)–conjugated goat antirabbit IgG antibodies (Chemicon, Temecula, CA) for 1 hour at room temperature. Negative controls consisted of slides in which only the second antibody was added. Slides were examined using a × 63 objective under a Zeiss LSM 410 confocal laser scanning system attached to the Zeiss Axiovert 135 M inverted microscope with × 63/1.2 C-Apochromat water immersion lens (Carl Zeiss, Thornwood, NY).

Data are expressed as mean ± SEM of at least 3 independent experiments. Parametric analysis (ANOVA, followed by Tukey-Kramer post hoc test) was used to compare the effects. In both cases,P <  .05 was considered statistically significant.

To evaluate whether human lung mast cells can be activated by MBP under fibroblast influence, human lung mast cells were cocultured for 4 days with human lung fibroblasts. Human lung mast cells cocultured with human lung fibroblasts released histamine in a concentration-dependent fashion with a maximum effect at 10 μM MBP (51.4% ± 5.5% histamine release; P < .001; n = 5; Figure1).

At the same range of MBP concentrations, human lung mast cells maintained in suspension in the presence of SCF (100 ng/mL) did not release significant levels of histamine compared with buffer-incubated cells (data not shown). The increased reactivity of cocultured human lung mast cells toward MBP cannot be related to a mast cell change in phenotype toward MC_TC. In fact, as assessed by confocal microscopy, after coculture the majority of human lung mast cells were still tryptase-containing rather than tryptase/chymase-containing cells (data not shown). This finding is in agreement with a previous study carried out by Dvorak et al.21 

The addition of either MBP (1 μM) or compound 48/80 (5 μg/mL) to CBMC cultures induced the release of relatively low percentages of histamine, that is, 16.1% ± 1.4% and 19.7% ± 1.9%, respectively (P < .05; n = 26; data not shown). Therefore, the modest effect of MBP and 48/80 on mast cell histamine release prompted us to search for culture conditions in which CBMC susceptibility to IgE-independent activation could be increased.10 CBMCs cocultured with 3T3 fibroblasts released 2 times more histamine (60.8% ± 6.7%) than CBMCs maintained in suspension (31.8% ± 3.6%) when activated by MBP (1 μM; P < .05; n = 5). CBMCs cocultured with 3T3 released even higher percentages of histamine when the MBP concentration was increased to 2 μM (28.6% ± 4.5% histamine release for CBMCs maintained in suspension versus 73.7% ± 5.6% for CBMCs cocultured with 3T3; P < .0001; n = 3). Similarly, compound 48/80 (5 μg/mL) induced a 2.9-fold increase in histamine release from CBMCs cocultured with 3T3 (P < .05; n = 5; Figure2A). In contrast to these effects, no differences in histamine release were obtained when CBMCs cocultured with 3T3 were activated immunologically by the anti-FcεRI-α chain antibody, 22E7 (2 μg/mL).

PGD₂ release was also enhanced following coculture with 3T3 fibroblasts. In fact, CBMCs from 2 different donors activated with MBP (1 μM) generated 7.7 ± 1.1 and 88.9 ± 39.8 ng PGD₂/10⁶ cells (P < .001), whereas the same CBMCs maintained in suspension produced only 1.0 ± 0.6 and 10.5 ± 0.9 ng PGD₂/10⁶cells, respectively (Figure 2B-C).

In contrast to PGD₂ production, CBMCs activated by MBP did not generate a significant amount of cys-LTs both in suspension and in coculture conditions (data not shown). In control experiments, no cys-LT or PGD₂ was released from 3T3 fibroblasts that had been incubated with 22E7 antibodies (2 μg/mL), 48/80 (5 μg/mL), or MBP (1 μM).

To assess whether a soluble fibroblast-derived factor might be responsible for the high mast cell susceptibility to IgE-independent activation, CBMCs were incubated with conditioned media obtained from CBMCs cocultured with 3T3 fibroblasts or were separated from 3T3 fibroblasts using Transwell membranes. Under both culture conditions, mast cell responsiveness to MBP (1 μM) was significantly lower than that of CBMCs cocultured directly on 3T3 (Figure3A). This suggests that a fibroblast membrane-associated factor rather than a soluble one is responsible for the fibroblast influences on CBMCs. The membrane-bound form of SCF has shown to partially mediate the effects of murine lung fibroblasts and human endothelial cells on murine bone marrow–derived mast cell responsiveness to IgE-dependent activation and human intestinal mast cell survival and proliferation, respectively.14,22Therefore, we investigated whether membrane SCF expressed on the fibroblasts during coculture is responsible for the increased susceptibility of human mast cells toward non-IgE–dependent activation. For this purpose, the fibroblast expression of the membrane form of SCF was modulated using antisense oligonucleotides for SCF. CBMCs cocultured with 3T3 and treated with antisense SCF oligonucleotides (0.1 pM) released 83.3% less histamine than sense-treated cells when activated by MBP (1 μM;P < .05; n = 3; Figure 3B). Antisense-treated cells were also completely inhibited in their production of PGD₂after stimulation with MBP (1 μM; data not shown). Similar results were obtained when CBMCs were cocultured with embryonic fibroblasts from Sl/Sl^d mutant mice that produce the soluble but not the membrane form of SCF (Figure 3C).23 

We next determined one of the possible mechanisms by which mast cells in coculture with fibroblasts increase their responsiveness to IgE-independent activation. It is known that IgE-independent secretagogues stimulate rat peritoneal mast cells by directly activating G_i3, a Ptx-sensitive G_i protein that leads to mast cell exocytosis and eicosanoid generation.24,25 Therefore, we investigated the expression of G_i3 protein in CBMCs in suspension and after coculture. As shown in Figure 4A, coculture of CBMCs with 3T3 fibroblasts resulted in a marked increase (6.2-fold) in the expression of the α subunit of G_i3 compared with CBMCs maintained in suspension as determined by immunoblotting. Preincubation with Ptx significantly inhibited histamine and PGD₂released by MBP-activated CBMCs (75.4% and 94.4% inhibition, respectively).

In confocal microscopy images, G_i3α expression appeared to be highly localized in the plasma membrane of the CBMCs cocultured with 3T3 fibroblasts (Figure 4B). Similar results of increased presence of G_i3α were obtained in human lung mast cells cocultured with human lung fibroblasts (data not shown). Immunoblotting analysis of CBMCs maintained in suspension did not reveal any expression in Gi_3α (Figure 4C). Freshly isolated rat peritoneal mast cells, as expected, constitutively expressed G_i3α24 (Figure 4D).

As assessed by immunoblotting and further densitometric analysis, CBMCs cocultured with 3T3 and treated with antisense for SCF expressed less G_i3α than the sense-treated cocultures (2.4-fold; Figure5A) and less G_i3α than the untreated cocultured mast cells (3.7-fold; Figure 4A). A similar decrease in G_i3α expression was observed in confocal microscopy images of antisense (Figure 5B) compared with sense-treated cocultures (Figure 5C). Furthermore, the same cocultures treated with antisense for SCF were unresponsive to MBP (Figure 3B).

In this work, we have demonstrated for the first time that highly purified human lung mast cells are responsive to the eosinophil mediator MBP when in coculture with fibroblasts. Coculture with fibroblasts is an in vitro culture system that more closely resembles in vivo conditions than the in vitro culture system usually used, that is, keeping the cells in suspension. Furthermore, CBMCs cocultured with fibroblasts were more responsive to MBP and compound 48/80 compared with the same cell population maintained in suspension.

It has been shown that incubation of rat peritoneal mast cells with native MBP and eosinophil cationic protein (ECP) but not eosinophil-derived neurotoxin (EDN) results in a concentration-dependent histamine release that requires both [Ca²⁺] and metabolic energy.26 In addition, we have recently demonstrated that IgE-desensitized rat peritoneal mast cells release histamine following incubation with MBP. The mechanism causing this release is similar to that of IgE-independent stimuli induced by cationic compounds such as compound 48/80 and substance P.4 However, human mast cell responsiveness toward eosinophil mediators remains less extensively investigated as yet. A previous study has shown that only partially purified human heart mast cells, but not human lung and skin mast cells, released histamine, tryptase, and PGD₂ when incubated with ECP and MBP.5 In the present study, we took into account that mast cell functionality/phenotype can be dramatically affected by the microenvironment.1 For example, murine bone marrow–derived mast cells that are mucosal-type mast cells, in coculture with 3T3 fibroblasts, changed their phenotype toward connective tissue mast cells, the rodent counterpart of MC_TC, and acquired responsiveness to IgE-independent stimuli in these culture conditions.10,11 Importantly, mast cells lost their responsiveness to IgE-independent activation when the cocultures are dispersed.27 This observation suggests that mast cell functionality or phenotype (or both) may be rapidly altered once the cells are isolated from the tissue influence. This might be in line with the lack of responsiveness of freshly isolated human lung mast cells toward IgE-independent activation. In the present study, we found that human lung mast cells respond to MBP after coculture with human lung fibroblasts, whereas freshly isolated and highly purified human lung mast cells are unresponsive to the same stimulus. Moreover, CBMCs that were only slightly responsive to MBP when challenged in suspension, on coculture with 3T3 fibroblasts became more responsive to this activation and released more histamine and PGD₂, but not cys-LTs. Similar results were found with another cationic secretagogue, compound 48/80.

The preferential production of PGD₂ over cys-LTs by mast cells when activated by IgE-independent stimuli can be attributed to the larger requirement for Ca²⁺ by the lipoxygenase than by the cyclooxygenase pathway to exert their enzymatic activities.28 In fact, it is known that the increase in intracellular Ca²⁺ concentration, after mast cell IgE-independent activation, is derived mainly from intracellular stores and is lower than that induced after IgE-dependent activation, derived from both extracellular and intracellular stores.29 In agreement with this, we observed that CBMCs cocultured with 3T3 fibroblasts produced and released more cys-LTs than PGD₂after IgE-dependent activation (130 ± 43.0 pg/mL cys-LTs compared with cells maintained in activation buffer that released only 20.0 ± 5.0 pg/mL cys-LTs). Therefore, the increase in PGD₂ generation during mast cell coculture with fibroblasts is not compensating a possible down-regulation of the signaling pathways that lead to cys-LT production.

Besides the low calcium requirements of PGD₂production,28 up-regulation in the cyclooxygenase (COX) expression may occur during mast cell coculture with fibroblasts. Indeed, it has been recently shown that murine bone marrow–derived mast cells in coculture with fibroblasts express higher levels of prostaglandin endoperoxide synthase mRNA and protein after stimulation with compound 48/80.10 A similar change in arachidonic acid profile due to an induction of leukotriene C₄ synthase expression has been recently reported in CBMCs exposed to IL-4.30 However, in preliminary experiments we could not detect COX-2 up-regulation in CBMCs cocultured with 3T3 fibroblasts (data not shown).

Some of the fibroblast influences on mast cells such as survival, proliferation, maturation, and IgE-dependent activation are known to be mediated by SCF.31-34 Furthermore, SCF induces eosinophil recruitment in the murine model of allergic airway inflammation by priming mast cells to synthesize higher levels of eotaxin, a strong eosinophil chemoattractant factor.35 This together with our observation of SCF produced by eosinophils,20 are strong evidence of the importance of SCF in allergic inflammation.

SCF is normally found in both soluble and membrane forms as a result of differential splicing and proteolytic cleavage.36 The differential effects of the soluble and membrane form of SCF on mast cells have not yet been extensively studied. However, the involvement of a fibroblast membrane-derived factor, rather than a soluble one, prompted us to consider SCF in its membrane form as the possible regulator of mast cell responsiveness to IgE-independent stimulus. Indeed, mast cells cocultured with fibroblasts in the presence of the soluble, but not the membrane, form of SCF released much less histamine and PGD₂ when incubated with MBP. In contrast, mast cell survival and adhesion to the fibroblasts in the same culture conditions were not affected by the absence of the membrane form of SCF (data not shown). Similarly, it has been shown that administration of exogenous soluble SCF to Sl/Sl^d mice increased mast cell numbers indicating differential effects of the 2 forms of SCF on mast cell responses.37 

SCF triggers its biologic effects by binding to its receptor, c-kit, a member of the type III receptor tyrosine kinase family.37-39 The existence of different isoforms of c-kit for soluble and membrane forms of SCF has not been reported. However, several works have provided evidence that the membrane forms of some growth factors are involved in signaling events distinct from those mediated by the diffusible forms even when acting on the same receptor. For example, growth and survival regulation have been attributed specifically to the membrane form of the heparin-binding epidermal growth factor (EGF)–like growth factor (proHB-EGF), rather than to the diffusible, processed HB-EGF isoform.40 Moreover, the membrane form of SCF has been shown to induce a prolonged activation of c-kit, inducing more efficient signals for the survival and differentiation of the factor-dependent myeloid cell line MO7e.41 In accordance with this, we observed that the membrane form of SCF up-regulates the expression of the α subunit of G_i3. This G_i3 overexpression may contribute to mast cell responsiveness to the IgE-independent activator, MBP. Indeed, similar correlation between G_i3 expression and responsiveness to compound 48/80 was found in quercetin-treated rat peritoneal mast cells.42 G_i3 is expressed both in the Golgi apparatus and in the plasma membrane. It is the plasma membrane-bound form of G_i3 that appears to facilitate regulated exocytosis because it is more accessible to interact with cationic secretagogues.24 In correlation with this, we found that G_i3α expression appeared to be highly localized in the plasma membrane of CBMCs. Besides this observation, it is possible that other changes in IgE-independent signaling can be induced during mast cell–fibroblast coculture as well, for example, a decrease in the rate of Gα guanosine triphosphatase (GTPase) activity that regulates G_i3-mediated exocytosis.43 

In conclusion, we demonstrated that human lung mast cells and human CBMCs can be activated by an allergen-independent eosinophil-derived stimulus. This event can feasibly occur during the late phase of the allergic process when mast cells encounter the infiltrated and activated eosinophils. Mast cell responsiveness to eosinophil mediators is specifically influenced by the membrane form of SCF.

Together these findings are important in the understanding of microenvironment influences on mast cell functionality that dictate either the resolution or the aggravation of a pathologic condition through the interaction with other inflammatory cells such as eosinophils or structural cells such as fibroblasts.

We wish to thank Prof Ronit Sagi-Eisenberg and Prof Hermona Soreq for helpful discussions. We also thank Dr Mark Tarshish for technical assistance with confocal microscopy images and Mrs Madelyn Segev for her excellent secretarial assistance.