Adhesion of human mast cells to extracellular matrix provides a co-stimulatory signal for cytokine production

Abstract

Engagement of integrin receptors during cell adhesion leads to changes in the morphology and the state of activation of cells. We therefore examined whether mast cell adhesion to extracellular matrix proteins affects the synthesis and release of various proinflammatory cytokines. Cells of the human mast cell line HMC-1 were added to fibronectin (FN)-, vitronectin (VN)- or, as a control, bovine serum albumin (BSA)-coated wells and were stimulated with phorbol 12-myristate 13-acetate (PMA) and/or calcium ionophore A23187 (ionophore). Cytokine production was evaluated using semiquantitative reverse transcription–polymerase chain reaction (RT–PCR) analysis of cell extracts and enzyme-linked immunosorbent assay (ELISA) analysis of cell supernatants. After a 4-hr incubation, mRNA expression of interleukin (IL)-8 (and weakly of IL-6) was up-regulated in matrix-adherent cells, with further increase in the presence of PMA and/or ionophore, compared with unstimulated cells. High-level de novo expression of IL-3 and of granulocyte–macrophage colony-stimulating factor (GM-CSF) was observed mainly in matrix-adherent cells. These changes were paralleled by the secretory pattern of HMC-1 cells after a 24-hr stimulation. Unstimulated cells adherent to FN or VN had already released small amounts of IL-8, and both VN- and FN-adherent cells produced, almost invariably, a higher level of cytokines than BSA-exposed cells after additional stimulation. These results show that mast cell adhesion to matrix proteins by itself has only selected and minor effects, but additional activation of mast cells by secretory stimuli causes significantly enhanced cytokine gene expression and secretion, suggesting that mast cells are far more active in their natural tissue environment than hitherto suggested from data in suspension cultures.

INTRODUCTION

The ability of mast cells to produce a variety of cytokines under appropriate conditions suggests that the cells can participate in immunological processes other than immunoglobulin E (IgE)-mediated hypersensitivity reactions. In support of this view, several recent reports have demonstrated the presence of mast cells and have provided evidence for their possible participation in more persistent, and even in chronic inflammatory and immunological, responses. During these processes, accumulation of mast cells in diseased tissue can vary markedly, depending on the prevailing type of inflammation.^1,2

Normally, mast cells are preferentially located adjacent to blood vessels, nerves or skin appendages in mucosal or connective tissue where they release their granule contents or secrete various cytokines upon stimulation.^3–5 The tissue-specific localization of mast cells is probably regulated by a certain cytokine milieu and by adhesion of the cells to extracellular matrix (ECM) components via specialized cell-surface receptors of the integrin family.⁶

In recent studies, we have shown that human mast cells adhere to the ECM proteins fibronectin (FN) and vitronectin (VN) via α₅β₁ and α_vβ₅ integrins, respectively.⁷ Evidence derived from in vitro studies indicates that engagement of integrin receptors transduces signals from the extracellular environment to the cytosol, leading to changes in the phenotype, movement and state of activation of the cells (reviewed in refs ^8,9). It is therefore likely that receptor-mediated contact of mast cells to ECM components may also influence their capacity to produce cytokines and consequently their ability to affect inflammation.

The aim of the present study was therefore to examine whether adhesion of the human mast cell line HMC-1 to the ECM components FN and VN might influence gene expression and protein secretion of the proinflammatory cytokines interleukin (IL)-3, IL-6, IL-8 and granulocyte–macrophage colony-stimulating factor (GM-CSF). These cytokines have been found to be expressed by stimulated HMC-1 cells.^4,5,10,11

MATERIALS AND METHODS

Cells and cell stimulation

HMC-1 cells, established from a leukaemia patient as an immature human mast cell line (kindly provided by J. H. Butterfield, Department of Allergic Diseases, Mayo Clinic, Rochester, Minneapolis, MN),¹² were maintained in Iscove’s medium, supplemented with 10% fetal bovine serum, 2 mm glutamine, antibiotics (all from Seromed, Berlin, Germany) and 10⁻⁵ m monothioglycerol (Sigma, Deisenhofen, Germany). Lymphokine-activated peripheral blood lymphocytes (LAK cells) (1×10⁶/ml) were cultured in RPMI-1640 medium (Seromed) containing 1000 U/ml recombinant human IL-2 (Sigma) for 5 days before harvest for RNA isolation. Adhesion assays were performed as published previously.⁷ Briefly, flat-bottomed 24-well plates were coated overnight at 4° with 200 μl of human FN (1 μg/ml) (Boehringer Mannheim, Mannheim, Germany) or human VN (7 μg/ml) (Gibco BRL, Gaithersburg, MD). Plates were rinsed with phosphate-buffered saline (PBS), and non-specific binding sites were blocked by incubation with 200 μl of PBS/3% bovine serum albumin (BSA; Sigma) per well for 1 hr at 37°. Plates were rinsed again, and a total of 1×10⁶ HMC-1 cells in 1 ml of serum-free Iscove’s medium/1% BSA were plated on each coated well in triplicate. Under these conditions, HMC-1 cells show ≈80% cell adhesion.⁷ As a control, cells were also added to wells that had been coated with PBS/3% BSA alone. To study the effect of activation on cell adhesion and cytokine production, phorbol 12-myristate 13-acetate (PMA) (Sigma; final concentration 42·5 nm) and/or calcium ionophore A23187 (Sigma; final concentration 5×10⁻⁷ m) were added to the cells immediately after plating. For RNA analyses, cells were pelleted in the plates and then lysed for extraction of total cellular RNA after 4 h of incubation. Accumulation of cytokines in supernatants of stimulated cells was determined after a 24-h incubation period. These incubation times and stimulation conditions have been found, by us, in the past and based on extensive kinetics, to be optimal for obtaining maximal cytokine mRNA and protein levels.^4,10,11

Reverse transcription–polymerase chain reaction analysis

RNA extraction, cDNA synthesis and semiquantitative reverse transcription–polymerase chain reaction (RT–PCR) were performed as previously described in detail.^10,11 Briefly, total cellular RNA (3 μg/ml), purified by the guanidium thiocyanate/caesium chloride method, was transcribed into cDNA using random priming and amplified during 30–35 cycles, using cytokine-specific oligonucleotide primers. For comparison of cytokine mRNA levels in different samples, cDNAs were first adjusted to equal concentrations of β-actin by the use of a β-actin control fragment (data not shown). To control for contamination, the PCR reaction mixture was amplified without cDNA or contained RNA that had not been reverse transcribed. Each PCR was repeated at least twice. Densitometric analysis of the bands obtained from amplification of cytokine mRNAs has been performed and results have been given as optical density values minus background values.

Cytokine assays

The amount of IL-6 protein in culture supernatants was determined by a sandwich enzyme-linked immunosorbent assay (ELISA), which was performed as described previously.¹¹ All other cytokines (IL-3, IL-8, GM-CSF) were determined using commercially available ELISA kits (Quantikine, R & D Systems GmbH, Wiesbaden, Germany), according to the instructions of the manufacturer.

Statistical analysis

Significance levels were determined using the two-sided Student’s t-test.

RESULTS

Matrix-dependent induction and augmentation of cytokine mRNA expression

To determine the influence of mast cell adhesion on cytokine gene expression, we compared the mRNA levels for IL-3, IL-6, IL-8 and GM-CSF of cells that had been incubated on FN- or VN-coated wells with those of cells that had been incubated on BSA-coated wells, using semiquantitative RT–PCR analysis. Figures 1 and 2 show the constitutive expression of mRNA for IL-8, and the lower constitutive expression of IL-6, on BSA-coated plates, with up-regulation of both in the presence of VN, and for IL-8 also in the presence of FN (Fig. 1, lanes 4, 8 and 12; Fig. 2).

Figure 1.

Cytokine gene expression by HMC-1 cells adhering to bovine serum albumin (BSA; lanes 4–7), fibronectin (FN; lanes 8–11) or vitronectin (VN; lanes 12–15), in serum-free medium (lanes 4, 8, 12), or co-stimulated with 42·5 nm of phorbol 12-myristate 13-acetate (PMA; lanes 5, 9, 13), 5×10⁻⁷ m of ionophore (lanes 6, 10, 14) or PMA and ionophore (lanes 7, 11, 15). Lane 1, 1-kb DNA ladder; lane 2, positive control (lymphocyte-activated killer cells); lane 3, negative control.

Figure 2.

Densitometric analysis of polymerase chain reaction (PCR) products after amplification of mRNA of HMC-1 cells adhering to bovine serum albumin (BSA), fibronectin (FN) or vitronectin (VN)-coated plates with interleukin (IL)-3, IL-6, IL-8 and granulocyte–macrophage colony-stimulating factor (GM-CSF) cytokine gene primers. Data show the optical density (OD) minus the background readings. 0=no stimulus added. Ionoph., ionophore.

Next, the relative contribution of matrix adhesion to cytokine production was studied in mast cells activated with PMA, ionophore, or with both stimuli combined. On BSA-coated wells, incubation of HMC-1 cells with PMA plus ionophore induced the highest mRNA levels for all cytokines except GM-CSF (Fig. 1, lane 7; Fig. 2). When the stimuli were studied individually, expression of IL-3 mRNA was induced de novo by the ionophore and not by PMA. IL-6 mRNA expression was up-regulated by the ionophore, and GM-CSF mRNA expression was induced mainly by PMA (Fig. 1, lanes 5 and 6). Neither stimulus alone increased IL-8 mRNA expression above the basal level.

Matrix adhesion significantly up-regulated the cytokine mRNA levels of HMC-1 cells activated with ionophore and/or PMA in a matrix- and cytokine-specific manner. Up-regulation of IL-8 gene expression after adhesion of unstimulated mast cells to FN was 282% and to VN 175% of the control values, whereas IL-6 gene expression increased only after adhesion to VN (193%), as was the case for GM-CSF (920%). No effects of either FN or VN were noted for IL-3 mRNA under these conditions (Fig. 1, lanes 8, 12; Fig. 2).

Compared with FN, adhesion to VN induced higher expression of IL-6 mRNA in PMA-stimulated cells, and of IL-8 following ionophore stimulation. FN was conversely more favourable for IL-8 mRNA production in unstimulated and PMA-stimulated cells and of IL-3 in ionophore-stimulated cells. On co-stimulation with ionophore and PMA, mRNA levels were, however, similar on either ECM protein, except for IL-3 where mRNA levels were lower on VN- than on FN-stimulated cells (lanes 9–11 and 13–15, Figs 1, 2).

Modulation of cytokine secretion by matrix proteins

To evaluate whether matrix adhesion might also result in modulated cytokine secretion of mast cells, the cytokine levels in supernatants from 24-hr stimulated mast cells were analysed using ELISAs. In agreement with the data on mRNA levels, ELISA data (Fig. 3) showed that adhesion to FN or VN by itself had no significant secretory effect, but that both ECM molecules function as co-stimulatory signals by enhancing cytokine release from HMC-1 cells activated with ionophore and/or PMA.

Figure 3.

Summary of cytokine production in HMC-1 cell supernatants. Cells adhered to bovine serum albumin (BSA), fibronectin (FN) or vitronectin (VN)-coated plates in the presence or absence of 42·5 nm of phorbol 12-myristate 13-acetate (PMA) and 5×10⁻⁷ m ionophore. Cell supernatants were harvested after 24 hr, and cytokine levels were assayed by enzyme-linked immunosorbent assay (ELISA). Values represent the means±SD of three independent experiments. 0=no stimulus added. Ionoph., ionophore. *P<0·001.

In cells exposed to both stimuli, adherence to FN resulted, however, in a more marked increase (300%) of IL-3 secretion than adherence to VN (172%), compared with cells seeded on BSA-coated wells (P<0·001). For IL-6, the effect of VN was generally more pronounced, with low-level induction of IL-6 secretion upon stimulation with either ionophore or PMA. In the presence of both stimuli, higher levels of IL-6 were secreted; furthermore, cultures adhering to VN secreted the highest levels of IL-6, with an increase above BSA-coated wells of 147% (P<0·001). For IL-8, adhesion to VN and FN resulted in a comparable increase of IL-8 secretion in the presence of either stimuli alone or in combination (277% and 262%, for FN and VN respectively, versus BSA in ionophore+PMA-activated HMC-1 cells, P<0·001) (Fig. 3). Enhancement of GM-CSF release by binding to FN or VN was also comparable in ionophore plus PMA-treated cultures (up to 371% in FN-coated wells, P<0·001), with a smaller effect in cultures treated with PMA or ionophore alone.

The role of integrins during FN- and VN-induced release of IL-8 was tested next by preincubating cells with specific adhesion-blocking anti-integrin monoclonal antibodies (mAbs) before seeding on FN (anti-β₁-integrin mAb: clone 4B4, dilution 1:50) or VN (anti-α_vβ₅-integrin mAb: clone P1F6, dilution 1:50) and stimulation with PMA plus ionophore. These mAbs have been shown by us to completely block HMC-1 adhesion to FN and VN.⁷ For both mAbs, a reduction of the increased release of IL-8, relative to levels of the BSA control, was demonstrated in both FN- and VN-cultured cells (data not shown).

DISCUSSION

The present study has examined for the first time the potential of human mast cells to modulate their cytokine response after binding to ECM proteins. The data show that adhesion of mast cells to ECM proteins significantly enhances cytokine gene expression and secretion following stimulation with an appropriate primary signal provided by the ionophore and/or PMA. Stimulation of mast cells by these signals was almost invariably necessary to elicit a secretory response, with marked effects for all cytokines, except IL-3, on exposure to the combined stimuli, and even with synergistic effects for IL-6 and GM-CSF. Adhesion of mast cells to ECM proteins in the absence of other stimuli either induced expression of the different proinflammatory cytokines in a highly variable manner (IL-6, IL-8) or had no effect on cytokine expression (IL-3, GM-CSF). Induction of IL-8 secretion up to levels of 100–900 pg/ml had already occurred after adhesion to FN or VN in the absence of ionophore or PMA. Although these quantities of IL-8 are low compared with those of PMA- and/or ionophore-stimulated cells, they might play a role in the maintenance of tissue homeostasis in the context of normal physiological processes.

Induction of cytokine secretion by matrix-adherent mast cells is probably integrin dependent. In previous studies, we demonstrated the exclusive engagement of α₅β₁ and α_vβ₅ by HMC-1 mast cells for adherence to FN and VN.⁷ Moreover, addition of mAbs to block adhesion completely abrogated the induction of IL-8 secretion over the basal level in this study. For all the other cytokines, significant enhancement of cytokine production by ECM-adhering mast cells occurred only after their activation by PMA and/or ionophore. Therefore, integrins act as co-stimulatory signals for cytokine transcription in mast cells.

Ionophore and/or PMA are apparently necessary for initiating gene transcription and to make mast cells susceptible to the additional stimulation during their contact with the ECM. Activation of signalling pathways associated with both calcium (ionophore) and protein kinase C (PMA) was obviously more efficient than activation of each pathway separately because, in the presence of both stimuli, induction of cytokine gene transcription was always found to be stronger.

Previously, enhancement, but not initiation, of inflammatory gene expression was also described following ligand-dependent engagement of CD11b/CD18 (Mac-1) integrins, for tissue factor (TF) or tumour necrosis factor (TNF)-α.¹³ Also in macrophages, integrin-mediated adhesion to FN acts as a strong GM-CSF mRNA inducer,¹⁴ while collagen binding induces high levels of TNF-α and decreased levels of colony stimulating factor-1.¹⁵ It has also been demonstrated that engagement of β₁ integrins, but not β₂ integrins, with specific mAbs as surrogate ligands can directly initiate the expression of several immediate-early genes in monocytes independent of adhesion.¹⁶ In an additional study, induction of TF gene transcription mediated by engagement of α₄ or β₁ has been shown to involve a cis-acting integrin-responsive element, which contained two binding sites for the transcription factor AP-1 and one site for κB-like transcription factors.¹⁷ It is thus conceivable that, as in monocytes, integrin-responsive elements enhance cytokine gene transcription also in mast cells, probably in co-operation with factors induced by ionophore and/or PMA stimulation. Although the use of phorbol ester and ionophore to stimulate mast cells is artificial and may not accurately reflect in vivo situations, both substances mimic mast cell activation via FcεRI-stimulation, generating an activation of protein kinase C and an elevation of [Ca²⁺]_i.

The present results suggest that synthesis of cytokines by mast cells may be part of the mechanism(s) involved in regulating inflammation, by attracting mast cells to tissue sites where they adhere and become activated, allowing the in situ proliferation and differentiation of the cells from their precursors that have migrated from the blood or more distant sites. Engagement of specific integrin receptors on immature mast cell precursors during adhesion to different substrates and subsequent transmigration through the endothelium has relevance, not only in determining the tissue-specific localization of mast cells. It may further influence the biological responsiveness of mature mast cells, retained within the connective tissue, by selectively enhancing the expression of mast cell-derived cytokines and evoke a specific immune and inflammatory response depending on the substrate they are adhering to.

Abbreviations

BSA

bovine serum albumin

ECM

extracellular matrix

FN

fibronectin

LAK cells

lymphokine-activated killer cells

mAb

monoclonal antibody

VN

vitronectin