SEARCH

SEARCH BY CITATION

Keywords:

  • Allergic inflammation;
  • CD48;
  • co-culture;
  • eosinophils;
  • mast cells

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Background

Mast cells (MCs) and eosinophils (Eos), the key effector cells in allergy, are abundantly co-localized particularly in the late and chronic stages of allergic inflammation. Recent evidence has outlined a specialized ‘allergic effector unit’ in which MCs and Eos communicate via both soluble mediators and physical contact. However, the functional impact of this bi-directional crosstalk on the cells' effector activities has not yet been revealed. We aimed to investigate whether MC/eosinophil interactions can influence the immediate and late activation phenotypes of these cells.

Methods

Human and murine MCs and Eos were co-cultured under various conditions for 1–2 h or 1–3 days, and in selected experiments cell–cell contact was blocked. Cell migration and mediator release were examined, and flow cytometry was applied to stain intracellular signaling molecules and surface receptors.

Results

Eosinophils enhanced basal MCs mediator release and co-stimulated IgE-activated MCs through physical contact involving CD48–2B4 interactions. Reciprocally, resting and IgE-stimulated MCs led to eosinophil migration and activation through a paracrine-dependent mechanism. Increased phosphorylation of activation-associated signaling molecules, and enhanced release of tumor necrosis factor α, was observed in long-term co-cultures. Eosinophils also showed enhanced expression of intercellular adhesion molecule 1, which depended on direct contact with MCs.

Conclusions

Our findings reveal a new role for MC/eosinophil interplay in augmenting short- and long-term activation in both cells, in a combined physical/paracrine manner. This enhanced functional activity may thus critically contribute to the perpetuation of the inflammatory response in allergic conditions.

For decades, mast cells (MCs) and eosinophils (Eos) have been regarded as the dominating cells in allergic inflammation [1-6]. In hypersensitivity reactions, these two cells are mostly recognized for their individual roles in orchestrating the acute and late phases of the response: Allergen-induced activation of tissue-dwelling MCs through FcεRI leads to their immediate degranulation and release of preformed and newly synthesized mediators, while Eos are subsequently recruited and activated, inducing successive inflammatory events and, together with the MCs, also mediating tissue remodeling [1-6].

It is well accepted that late/chronic allergy involves both MCs and Eos, yet less is defined about the manner by which these two cells co-function to perpetuate the response. Considering their abundance and persistence in advanced inflammatory stages, MCs and Eos may cross-talk in a paracrine manner [7, 8]. Moreover, we recently uncovered a novel specialized mode of communication between the two cells, via physical contact [9, 10]: This interaction is stable, tight, and physiologically relevant, as high rates of MC-Eos coupling were detected in human and murine allergic disorders [9]. Cell–cell contact, together with soluble growth factors, was found to mediate an MC-induced increase in Eos viability [9].

Similarly, it is conceivable that the ensemble of soluble/physical interactions between MCs and Eos, termed the Allergic Effector Unit (AEU) [9], affects also their effector activities by facilitating bi-directional exchange of regulatory information and transfer of signals. Several plausible pathways could enable the cells to modulate each other's functions. Early studies have shown that mediators of MCs affect Eos, and vice versa [7, 8]: Eos-derived major basic protein (MBP) can activate MCs in a non-IgE-dependent manner [11-13], and Eos peroxidase (EPO) is taken up by MCs [10, 14]. On the other end, tryptase induces Eos influx and release of EPO, interleukin (IL)-6, and IL-8 [15-17]. Eosinophils-stimulating properties are attributed to other MC products, that is, platelet activating factor (PAF) [18] and histamine [19]. Mast cells might also promote Eos migration by releasing histamine, prostaglandin (PG) D2, and eotaxin that can activate Eos via the H4 receptor, the chemoattractant receptor-homologous molecule on Th2 (CRTH2), and the chemokine receptor CCR3, respectively [19-22]. Effector functions may also be regulated by direct MC-Eos interactions: Cell surface molecules implicated in the MC-Eos contact mechanism (i.e., the CD2-family molecule CD48 and its high-affinity ligand 2B4, the adhesion molecules DNAM-1, and Nectin-2) [9, 23-25] convey stimulatory signals in these cells [9, 23, 26, 27]. In support of this, the activity of MCs, as well as Eos, is altered following physical contact with other immune cells [28-31].

In view of these diverse cell–cell communication mechanisms, the net outcome of a complex MC-Eos cross-talk in a physiological setting is still unknown. It remains to be seen whether such MC-Eos contacts can indeed enforce meaningful changes on cellular functionality and partake in shaping the allergic response in which they occur. Here, we explored these questions using a customized whole-cell co-culture system employing human/murine-derived MCs and Eos. Both cell types showed short- and long-term functional changes. These findings newly demonstrate that the MC-Eos AEU augments cellular effector functions and suggest that it has a role in intensifying and prolonging the late/chronic stages of allergy.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Cells and assays

Human cord blood-derived MCs (CBMCs), human peripheral blood Eos (pbEos), murine bone marrow MCs (BMMCs), and murine bone marrow Eos (BMEos) were prepared as described previously in [13, 23, 32-34] (Data S1). Mast cells-Eos conjugates were observed in 1 h multispectral-imaging flow cytometry (MIFC) experiments. In short-term co-culture activation experiments (ca. 1 h), release of β-hexosaminidase (β-hex), tryptase, and EPO was assessed by colorimetric assays, while phosphorylated intracellular proteins and lysosomal-associated membrane protein-1 (LAMP-1) were measured by flow cytometry. Eosinophils chemotaxis was examined in 1.5–3 h co-cultures by transmigration assays. In long-term co-cultures (1–3 days), cytokine release was analyzed by enzyme-linked immunosorbent assay (ELISA), and receptor expression and intracellular signaling were examined by flow cytometry. See Data S1, for detailed procedures.

Co-culture conditions

In most experiments, MC-Eos mixtures were predominantly at the 1 : 1 ratio (unless indicated otherwise). Control monocultures of either MCs or Eos were always set at the exact conditions as the respective co-cultures, to maintain equivalent concentrations and to allow for proper comparison between samples. In selected experiments, MCs and Eos were physically separated by 0.4-μm transwell inserts (Greiner Bio-One, Frickenhausen, Germany). Other co-culture settings were optimized for every particular assay (Data S1).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Eosinophils undergo chemotaxis toward resting and activated MCs

As MCs release factors that may induce Eos chemotaxis [20-22], we first examined whether, and to what extent, CBMCs would attract pbEos. In a transmigration assay, pbEos migrated significantly more toward IgE-activated CBMCs (a mean 45% of total cell migrating) than toward control medium (17%) or toward resting CBMCs (29%), indicating a chemotaxis effect rather than chemokinesis (Fig. 1A). Interestingly, the migration index in the MC-activated co-culture (showing 2.6-fold higher chemotaxis) was only slightly lower than that obtained by the Eos chemoattractant eotaxin-1, in which a 3.5-fold migration (60%) was found (Fig. 1A). Similar migration rates (47%) were detected also when CBMCs were stimulated through a non-IgE-dependent pathway, using compound 48/80 (Fig. 1A). In murine co-cultures consisting of C57BL/6 (B6)-derived cells, BMEos chemotaxis to dinitrophenyl (DNP)-activated BMMCs (pre-sensitized with IgE anti-DNP) was also significantly higher (69%), as compared to resting BMMCs (17%) or control wells (6%), and resembled migration levels induced by mouse eotaxin-1 (74%; Fig. 1B).

image

Figure 1. Eosinophil chemotaxis toward resting/activated mast cells. (A) Transmigration of peripheral blood eosinophils (pbEos) toward resting, IgE- or compound 48/80-activated cord blood-derived mast cells (CBMCs). (B) Transmigration of bone marrow eosinophils (BMEos) toward resting or IgE dinitrophenyl (DNP)-activated bone marrow mast cells (BMMCs). In both experiments, controls contain medium alone, with DNP, or with human/mouse eotaxin-1. Eos/MC ratios were 1 : 2. Chemotaxis is expressed as a migration index (= 3, *< 0.05, **< 0.005). NA-not activated.

Download figure to PowerPoint

Eosinophils activate MCs in human/mouse co-cultures

Next, we explored the possibility that MC-Eos interactions [9, 10] affect the immediate functional capabilities of these cells. The activation phenotype of MCs under Eos regulation was first tested: CBMCs were mixed with pbEos at various cell/cell ratios (1 : 0.1–1 : 10) for 1 h, and their degranulation was assessed by β-hex release. Peripheral blood Eos increased the basal CBMC release by approximately 5%, particularly when Eos numbers equaled, or were more than, those of MCs (Fig. 2A). IgE-activated CBMCs co-cultured with increasing amounts of pbEos also demonstrated stronger degranulation, approximately 15% higher than that of IgE-stimulated CBMCs in monoculture (Fig. 2A). This co-stimulatory effect peaked at the 1 : 1 co-culture, yet was roughly 10% lower in the high MC/Eos ratios. This significant reduction indicates potential inhibiting processes taking place in Eos-dominating samples. The pbEos-induced boost in CBMC activation was evident also by elevated tryptase release in baseline and IgE-stimulated settings (1 : 1 co-cultures, Fig. 2B). Neither β-hex nor tryptase were detectable in samples containing Eos alone.

image

Figure 2. Eosinophil-induced activation of mast cells. (A) Relative β-hex release from resting/IgE-activated cord blood-derived mast cells (CBMCs) in 1 h co-cultures with peripheral blood eosinophils (pbEos) at different ratios (n = 5, *< 0.05, different donors shown in colored lines, means-Ο). (B) Tryptase release from resting/IgE-activated CBMCs in 1 h co-cultures with pbEos at a 1 : 1 ratio (n = 5, *< 0.05). (C) Phosphorylated molecules in co-cultured CBMCs (MFI fold-increase, n = 3, *< 0.05). Tryptase release (D) and lysosomal-associated membrane protein (LAMP)-1 expression (ΔMFI) (E) in resting/activated bone marrow mast cells (BMMCs) co-cultured (1 : 1) with bone marrow eosinophils (BMEos) (n = 3, *< 0.05). No β-hex/tryptase was released by Eos. Data shown vs. control monocultures. NA-not activated.

Download figure to PowerPoint

In view of this MC degranulation, we hypothesized that signaling molecules typically associated with the early phases of MC responses to stimuli [35] might be also upregulated in our co-cultures. To isolate the Eos contribution to MCs in the absence of IgE stimuli (Fig. 2A), we measured intracellular phosphorylation events solely in the resting CBMC-pbEos co-cultures (1 : 1), for up to 2 h following co-culture onset. No signaling events could be found at any of the early time points (Fig. 2C). However, in 2 h co-cultures, we detected a strong increase in phosphorylation of the cytoplasmic tyrosine kinase Syk, as well as one of its target proteins, linker of activated T cells (LAT) (as evident by a rise in their MFI levels over monocultured cells), suggestive of belated signaling events. Changes in phosphorylation of the Lyn kinase were overall insignificant.

To examine whether these Eos-induced changes in MC activation levels occur also in the mouse system, B6 BMMCs and BMEos were similarly co-cultured. Cell–cell interactions at acceptable frequencies were confirmed (Data S2 and Fig. S1). In 1 h co-cultures with BMEos (1 : 1), BMMCs indeed displayed augmented release of tryptase (Fig. 2D) and β-hex (data not shown). This higher degranulation of co-cultured BMMC was observed in resting conditions, as well as under IgE activation by suboptimal (<30 ng/ml) DNP concentrations (Fig. 2D), but not when BMMC were fully stimulated by high (>30 ng/ml) DNP levels (data not shown). Tryptase was undetectable in BMEos monocultures. BALB/c-derived co-cultures showed similar results (data not shown). Lysosomal-associated membrane protein-1, a molecule associated with BMMC activation [36], was increased in B6- or BALB/c-derived BMMCs already within 30 min of co-culture with BMEos (1 : 1) in resting conditions (Fig. 2E), supporting heightened degranulation in BMEos-stimulated BMMCs.

Eosinophils-induced MC activation requires cell–cell contact via CD48/2B4

We previously identified the interaction between CD48 (on MCs) and 2B4 (on Eos) as one pathway mediating direct MC-Eos contact and modulation of Eos viability in the human AEU [9]. To understand whether the CD48-2B4 binding plays a role also in Eos-induced activation of MCs, short-term 1 h co-cultures (1 : 1 ratio) were carried out in receptor-blocking conditions. 2B4-neutralized PbEos lost their ability to augment IgE-activated CBMC degranulation (Fig. 3). Similar Ab neutralization of intercellular adhesion molecule (ICAM)-1 and Nectin-2 [two other Eos molecules that can bind MC receptors lymphocyte function-associated antigen (LFA)-1 and DNAM-1 [23, 24], respectively] also significantly reduced the activation of CBMCs in these co-cultures, albeit to a lower extent (Fig. 3). Blockade of the CD48/2B4 pathway similarly inhibited the BMEos-induced activation of BMMC in the murine co-cultures (see Data S3 and Fig. S2A,B).

image

Figure 3. Involvement of the CD48-2B4 axis in eosinophil activation of mast cells. Relative tryptase release from resting/IgE-activated cord blood-derived mast cells (CBMCs) in 1 : 1 co-culture with peripheral blood eosinophils (pbEos), neutralized for 2B4, intercellular adhesion molecule (ICAM)-1, or Nectin-2 (n = 3, *< 0.05). Control samples contain monocultures or co-cultures with isotype-treated cells. NA-not activated.

Download figure to PowerPoint

Eosinophils are activated by MCs in human/murine co-cultures independently of CD48-2B4-contact

We next tested whether, in a reciprocal pathway, MC can activate Eos. Cord blood-derived MCs (1 h, 1 : 1 ratio) stimulated pbEos to release ca. three-fold more EPO, that is, an estimated 800 ng/ml (47% Eos degranulation) in co-cultures vs. the 317 ng/ml (18% basal release) in control monocultures (Fig. 4A). We observed this effect also across diverse cell/cell ratios [10]. To assess whether the CD48-2B4 ligation mediates this increase, CD48 was preneutralized on CBMC. In contrast to the Eos co-stimulation of MCs (above), CD48 blockade did not impair CBMC activation of pbEos (Fig. 4A). Blockade of LFA-1 on CBMC was likewise ineffectual (Fig. 4A). These data indicate a lack of involvement of CD48/2B4 interactions in MC-induced Eos degranulation.

image

Figure 4. Mast cell-induced eosinophil activation independently of CD48/2B4. Eosinophil peroxidase (EPO) secretion from peripheral blood eosinophils (pbEos) co-cultured (1 : 1) with CD48/lymphocyte function-associated antigen (LFA)-1-neutralized cord blood-derived mast cells (CBMCs) (A), B6 bone marrow eosinophils (BMEos) co-cultured with bone marrow mast cells (BMMCs) at indicated ratios (B), and BMEos co-cultured (1 : 1) with strain-matched resting/IgE-activated BMMCs (C). (D) Lysosomal-associated membrane protein (LAMP)-1 expression (ΔMFI) in (1 : 1) co-cultured BMEos. Data shown vs. untreated/platelet activating factor (PAF)-activated cells, or isotype-treated co-cultures (n = 3, *< 0.05). NA-not activated. ns- not significant.

Download figure to PowerPoint

Higher EPO activity was also detected in BMEos co-cultured with BMMC (1 : 1, 1 h) as compared to BMEos alone (Fig. 4B, B6-derived cells). This activity (0.267 OD) corresponded to 660 ng/ml of EPO in co-culture supernatants (i.e., 39% Eos degranulation), an eight-fold increase over monocultured BMEos (0.105 OD; 4% Eos spontaneous release; 75 ng/ml EPO). Bone marrow MCs considerably stimulated BMEos (10% degranulation) even when reducing their amounts up to a quarter of BMEos numbers (Fig. 4B, 1 : 0.25 cell ratios). The MC-induced effect, however, was not further increased by IgE-stimulation: In 1 : 1 ratios, DNP-activated BMMCs (either B6- or BALB/c-derived) stimulated BMEos similarly to resting BMMCs (Fig. 4C). Of note, BMMC-induced BMEos stimulation appeared moderate compared to the potent PAF-induced stimulation that resulted in 82–96% degranulation (Fig. 4C). LAMP-1 upregulation verified that BMEos underwent degranulation (Fig. 4D). Of note, as in the human co-culture, CD48–2B4 contact was not implicated in BMMC activation of BMEos (Data S3 and Fig. S2C).

Mast cells and Eos maintain an activated phenotype in long-term co-cultures

As the AEU is long-lived [9], we sought to examine the long-term consequences of MC-Eos contact. An activated phenotype persisted even in 3-day co-cultures (1 : 1), as shown by elevated tumor necrosis factor (TNF) α levels in co-culture supernatants (greater than eight-fold increase, Fig. 5). No other tested cytokine (IL-6, -8, -10) was modulated in co-culture (data not shown). The increased TNFα secretion, likely by both cells [2, 3], required physical cell–cell contact: Transwell-separated co-cultures yielded only two-fold higher TNFα levels, as observed only in samples where MCs were placed above Eos (Fig. 5).

image

Figure 5. Sustained tumor necrosis factor-α (TNFα) release and activation of mast cells/eosinophils in long-term co-cultures. TNFα release in full-contact vs. transwell (tw) 3-day cord blood-derived mast cell (CBMC)/peripheral blood eosinophils (pbEos) co-cultures (1 : 1). Data are shown vs. monocultured cells (n = 3, *< 0.05).

Download figure to PowerPoint

We also explored surface receptor expression in long-term AEU: No changes were observed in AEU-mediating molecules (CD48, 2B4, DNAM-1, Nectin-2) following a 2-day co-culture (Fig. 6A). Of the screened adhesion molecules, ICAM-1 expression was strongly increased in co-cultured pbEos (Fig. 6A). This CBMC-stimulated elevation of pbEos ICAM-1 was apparent already after 1 day and increased in a time-dependent fashion throughout the 3-day co-culture (Fig. 6B). Interestingly, this effect required direct cell–cell contact because ICAM-1 was unaltered in pbEos within partial transwell-containing co-cultures (Fig. 6C, 48 h). These findings collectively show that long-term AEU interactions sustain MCs and Eos at a high activation level.

image

Figure 6. Intercellular adhesion molecule (ICAM)-1 expression on eosinophils in long-term mast cell containing co-cultures. (A) Expression of cell surface molecules in co-cultured cord blood-derived MCs (CBMCs) and peripheral blood eosinophils (pbEos) (2 days, 1 : 1, representative of n = 4). (B) ICAM-1 expression (ΔMFI) in 1–3 day co-cultured pbEos (n = 3, *< 0.05). (C) ICAM-1 expression in pbEos in full-contact vs. transwell co-cultures (2 days, 1 : 1, representative of n = 4). Data are shown vs. monocultured cells, and over isotype-staining controls.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

In this work, we have begun to define the outcomes of the AEU [9] on MC/Eos effector activities, gaining initial insight into the ability of this unique interface to shape the functions of these key cells in an advanced inflammatory response. Although soluble MC/Eos communication has long been suggested [7, 8], hitherto the evidence for such interactions was largely indirect, as most studies investigated the effect of purified MC/Eos granule constituents, or cell lysates/sonicates, in single-cell cultures. Thus, our findings are the first demonstration that MCs modulate Eos effector functions, and vice versa, using a whole-cell co-culture strategy in both human/mouse systems.

Our observation that Eos migrate toward activated MCs emulates the signals stemming from MCs immediately after allergen exposure, igniting Eos tissue influx [2, 3]. Moreover, augmented degranulation of co-cultured MCs and Eos was observed in cell/cell ratios that are physiologically feasible in acute or progressive inflammatory states: Eos activated MCs already at a 1 : 1 ratio, characteristic of initial Eos infiltration states, while MCs induced Eos degranulation even in an Eos-dominating setting (1 : 0.25 ratio), typical in later stages. However, despite the fact that MCs can activate Eos (and vice versa) by a number of distinct mediators [7, 8], our data do not show overt activation levels; this suggests that the different individual pathways may not necessarily synergize in the ‘whole-cell’ context. Indeed, this is a very reasonable finding: It is plausible that the cells’ mediators, when present in lower, more physiologically relevant concentrations than those used in the purified systems in the past, actually exert a lower activation effect. Another possible explanation for the subtle readouts is that Eos proteolytic factors (also likely released in co-culture) may have degraded MC mediators in the whole-cell system, and vice versa. Finally, we cannot exclude the possibility that inhibition of the cells also occurs, because MC activation was slightly reduced when Eos were ten-fold more frequent. Yet, the collective bi-cellular activation in our whole-cell setting is still very notable.

Interestingly, the activation phenotypes were boosted by the AEU under both resting and IgE-activated conditions. Although pbEos activation of resting CBMCs was subtle as compared to IgE stimulation, pbEos could still further activate IgE-stimulated CBMCs. The contribution of BMEos to BMMC degranulation under IgE-stimulation was more pronounced, because partial IgE-activation by suboptimal DNP was employed in these settings. Eosinophils signals could be additive to IgE stimuli, especially because their mediators (i.e., MBP) are known MC stimulants [13]. Eosinophils could also reduce the threshold of MC responsiveness to IgE, that is, conveying co-stimulatory signals integrating into IgE-mediated pathways. Indeed, phosphorylation events (heightened Syk and LAT) in the resting MCs under co-culture suggest that Eos alone ignite activation-associated signaling, albeit perhaps slower and milder than potent IgE stimuli. Taken together, bi-directional MC-Eos activation is likely of complex nature, potentially spanning across a wide time window of inflammation and varying depending on the allergen presence and cellular density. This may allow to maintain a continuously stimulated response throughout the dynamic allergic response.

The CD2 molecules CD48 and 2B4 contribute to immediate activation in MC-Eos co-cultures. Both molecules participate in stimulatory/co-stimulatory signaling in lymphocytes and natural killer (NK) cells [25, 37-40]. CD48, a murine asthma signature gene [25, 41] and a potential allergy biomarker in humans [25, 42], is a glycophosphatidylinositol-anchored molecule co-stimulating IgE-dependent secretion of MC inflammatory cytokines (Levi-Schaffer, unpublished data) and mediating MC bacterial uptake [43, 44]. Bone marrow MCs lacking glycophosphatidylinositol molecules possess weakened FcεRI signaling and impaired degranulation under IgE stimuli [45]. Our data provide further evidence for a co-activatory role for CD48, because 2B4-blocked Eos (that could not ligate CD48) were unable to augment IgE-stimulated human/mouse MCs. In view of the MC-Eos conjugation frequencies in our co-culture snapshots [ca. 1–5%, [9]], a multiphase effect may occur: The MC-Eos couples, activated via 2B4–CD48 contact, degranulate and release mediators, which may then induce further bi-directional activation of other cells which have not yet interacted with each other.

In contrast, the opposite route of MC-induced Eos degranulation did not need CD48–2B4 contact (as CD48-blocked MC properly activated Eos). This was an unexpected finding, because 2B4 ligation has been shown to provoke human pbEos degranulation [27]. This discrepancy might be explained by the fact that 2B4 stimulation in the early study was achieved by mAb cross-linking, while the ligand in our system was MC-expressed CD48, the latter possibly less potent. Alternatively, it may be that CD48-2B4 signaling does contribute to MC activation of Eos, yet other soluble/physical mechanisms compensate for its absence in the neutralized co-cultures, ultimately activating Eos despite the blockade. Notwithstanding, a working CD48-2B4 axis is probably important for other late Eos functions: MC induction of Eos survival [9] is partially attributed to 2B4-ligation. Our finding that CD48/2B4 expression is unhindered in long-term co-cultured MCs and Eos also supports this. Collectively, this demonstrates a prominent, yet nonsymmetrical, role for the CD48-2B4 pair in boosting immediate/chronic functions within interplaying MCs and Eos. Specific targeting of the AEU by CD48/2B4 inhibitors may thus be of benefit in allergy.

Other pathways may contribute to MC-Eos bi-directional activation. Mast cells mediators can stimulate Eos, and vice versa [7, 8]. Leukotrienes (LTs) produced by activated MCs/Eos may also interact with the cells, as both express LT receptors [1, 3]. Beyond the CD48-2B4 axis, other surface adhesion molecules and/or integrins with capacity to boost and extend immunological synapse signals could also partake in the MC-Eos contact-dependent interplay: The binding of MCs DNAM-1 to Eos Nectin-2, for example, has already been implicated in Eos-augmented degranulation of MCs [23], as observed also herein. A similar finding surfaced concerning the ICAM-1-involved pathway, as its blockade on pbEos reduced their ability to co-stimulate IgE-activated CBMCs. Bearing in mind that ICAM-1 augments cytokine-induced activation in Eos [46, 47], it is possible that ICAM-1-neutralized pbEos underwent poorer activation in co-culture, which in turn limited their capability to activate MCs. In this context, it is worth mentioning that ICAM-1/LFA-1 interactions also facilitate T-cell induction of MC degranulation in comparable IgE-involved settings [30, 48, 49].

As MCs and Eos persist in the inflamed tissue for significant periods, during which late-phase/chronic responses surface, AEU long-term phenomena also warranted our investigation. As expected, and in continuance to the boosted activation phenotypes observed in the short-term, late co-cultures demonstrated persisting pro-inflammatory effects. This involved increased TNFα secretion, likely released from both cells [1, 3], in a mechanism necessitating physical MC-Eos binding. Interestingly, TNFα is largely involved in MC induction of recruitment, proliferation and activation of lymphocytes, both in vitro and in allergic airway inflammation [50, 51]. Thus, by raising TNFα levels, the AEU might have an active part in shaping the characteristic Th2-profile in allergy.

In long-term co-cultures, MCs also raised ICAM-1 expression on Eos. Similar rising levels of the Eos adhesion molecule CD11b were seen following interaction with fibroblasts [52] or endothelial cells [53]. Interestingly, here, the MC-augmented Eos ICAM-1 effect required cell–cell contact, yet the mechanism could entail also paracrine communication between the cells: As histamine elevates ICAM-1 on Eos [19], immediate MC degranulation events in co-culture may contribute to this effect. Notwithstanding, TNFα, released in the long-term MC-Eos co-cultures, can also increase Eos ICAM-1 [54]. Communication via this proinflammatory cytokine likely occurs downstream of the critical cell–cell interactions, because transwell MC-Eos co-cultures failed to produce high TNFα concentrations. Moreover, considering that ICAM-1 signaling is linked to prolonged Eos survival [47], the rise in this molecule's levels on co-cultured Eos may be an additional mechanism underlying the MC-induced maintenance of these cells in the long term [9]. Concomitantly, this elevated Eos ICAM-1 expression, in tandem with CD48-2B4 binding, could serve to fortify the MC-Eos adhesion within late inflammation phases [as in the case of NK adhesion [55, 56]]. Altogether, these mechanisms could underlie the continuous occupation of active Eos within the inflamed tissue.

The above-described pathways and others, whether via soluble or physical factors, presumably integrate to additively or synergistically enhance MC/Eos activation, forming a ‘vicious cycle’. Thus, our results support the notion of a unique self-perpetuating MC/Eos AEU contributing to the persistence of allergic reactions. We speculate that, compared to potent allergen-induced activation, the AEU has a more subtle stimulatory ability, fine-tuning effector activities of MCs/Eos in the heat of active inflammation but also during late/chronic phases in which both cells are abundant. Co-stimulatory signals reduce the critical threshold of cellular activation in the immunological synapse [57-59]. By analogy, co-stimulation within the AEU may be exactly what is necessary for an activatory outcome, particularly when allergen stimuli are lacking, or when MCs are temporarily desensitized to allergens. As we have mentioned, inhibiting outcomes are also plausible, as MCs and Eos have inhibitory surface molecules [60] and immunoregulatory properties [1]. If such negative signals exist in the AEU, they may be initially overpowered by stimulatory pathways yet come to play later within allergic resolution [61, 62]. Future studies should reveal the AEU effects on MC/Eos responsiveness in diverse conditions.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

We thank the Levi-Schaffer lab members for helpful discussions. This work was supported by the Israel Science Foundation (grant 213/05), the MAARS EU 7th framework (grant no. HEALTH-F2-2011-261366), the Aimwell Charitable Trust (London, UK), and the Adolph and Klara Brettler Center for Research in Molecular Pharmacology and Therapeutics at the Hebrew University of Jerusalem.

Author contributions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

M.E. designed the research, performed all experiments, analyzed data, and wrote the manuscript; I.B. performed activation experiments; A.H.N.B.E helped to assemble the co-culture system; D.M. provided cord blood samples; and F.L.S. designed the research, received the grant supports, provided overall supervision, analyzed the data, and edited the manuscript.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Author contributions
  8. Conflict of interest
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
all12059-sup-0001-DataS1-S3.docWord document158K

Data S1. Supporting Methods.

Data S2. Visualization of the contact between BMMCs and BMEos and measurement of their interaction frequencies in vitro.

Data S3. Involvement of CD48-2B4 signaling in BMMC/BMEos bi-directional stimulation.

all12059-sup-0002-FigureS1.tifimage/tif1577KFigure S1. Physical contact between mouse-derived MCs and Eos in vitro.
all12059-sup-0003-FigureS2.tifimage/tif1429KFigure S2. Receptor expression in mature BMMC and BMEos cultures.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.