Pentacyclic triterpenoid ursolic acid interferes with mast cell activation via a lipid-centric mechanism affecting FcεRI signalosome functions

Pentacyclic triterpenoids, including ursolic acid (UA), are bioactive compounds with multiple biological activities involving anti-inflammatory effects. However, the mode of their action on mast cells, key players in the early stages of allergic inflammation, and underlying molecular mechanisms remain enigmatic. To better understand the effect of UA on mast cell signaling, here we examined the consequences of short-term treatment of mouse bone marrow-derived mast cells with UA. Using IgE-sensitized and antigen- or thapsigargin-activated cells, we found that 15 min exposure to UA inhibited high affinity IgE receptor (FcεRI)–mediated degranulation, calcium response, and extracellular calcium uptake. We also found that UA inhibited migration of mouse bone marrow-derived mast cells toward antigen but not toward prostaglandin E2 and stem cell factor. Compared to control antigen-activated cells, UA enhanced the production of tumor necrosis factor-α at the mRNA and protein levels. However, secretion of this cytokine was inhibited. Further analysis showed that UA enhanced tyrosine phosphorylation of the SYK kinase and several other proteins involved in the early stages of FcεRI signaling, even in the absence of antigen activation, but inhibited or reduced their further phosphorylation at later stages. In addition, we show that UA induced changes in the properties of detergent-resistant plasma membrane microdomains and reduced antibody-mediated clustering of the FcεRI and glycosylphosphatidylinositol-anchored protein Thy-1. Finally, UA inhibited mobility of the FcεRI and cholesterol. These combined data suggest that UA exerts its effects, at least in part, via lipid-centric plasma membrane perturbations, hence affecting the functions of the FcεRI signalosome.

Pentacyclic triterpenoids, including ursolic acid (UA), are bioactive compounds with multiple biological activities involving anti-inflammatory effects. However, the mode of their action on mast cells, key players in the early stages of allergic inflammation, and underlying molecular mechanisms remain enigmatic. To better understand the effect of UA on mast cell signaling, here we examined the consequences of short-term treatment of mouse bone marrow-derived mast cells with UA. Using IgE-sensitized and antigen-or thapsigargin-activated cells, we found that 15 min exposure to UA inhibited high affinity IgE receptor (FcεRI)-mediated degranulation, calcium response, and extracellular calcium uptake. We also found that UA inhibited migration of mouse bone marrow-derived mast cells toward antigen but not toward prostaglandin E 2 and stem cell factor. Compared to control antigen-activated cells, UA enhanced the production of tumor necrosis factor-α at the mRNA and protein levels. However, secretion of this cytokine was inhibited. Further analysis showed that UA enhanced tyrosine phosphorylation of the SYK kinase and several other proteins involved in the early stages of FcεRI signaling, even in the absence of antigen activation, but inhibited or reduced their further phosphorylation at later stages. In addition, we show that UA induced changes in the properties of detergent-resistant plasma membrane microdomains and reduced antibody-mediated clustering of the FcεRI and glycosylphosphatidylinositol-anchored protein Thy-1. Finally, UA inhibited mobility of the FcεRI and cholesterol. These combined data suggest that UA exerts its effects, at least in part, via lipid-centric plasma membrane perturbations, hence affecting the functions of the FcεRI signalosome.
Antigen-triggered activation of mast cells followed by the release of various inflammatory mediators, such as histamine, chemokines, cytokines, and products of arachidonic acid metabolism, are important events initiating allergic inflammation. In this process, the high affinity IgE receptor (FcεRI) complex localized at the plasma membrane plays a key role. FcεRI is a tetramer composed of the α subunit, which binds IgE, two γ subunits, which generate the downstream signal through the process initiated by tyrosine phosphorylation, and one β subunit, which amplifies the activation signals. FcεRIs, various enzymes, ion channels, and transmembrane adapter proteins involved in the FcεRI signaling form a functional complex embedded within the phospholipid bilayer of the plasma membrane, called the FcεRI signalosome. The FcεRI signalosome is a highly dynamic complex that interacts with and is regulated by various plasma membrane and cytoplasmic components and plays a crucial role in propagating the signal from FcεRI into the cytoplasm and the nucleus. This leads to degranulation of preformed secretory granules and de novo production and release of cytokines and other inflammatory mediators (1)(2)(3)(4)(5).
Recognition of the critical role of mast cells in allergic inflammation was the driving force in the search for new compounds that prevent mast cell activation, degranulation, and production of inflammatory mediators (3,6). One natural compound with anti-inflammatory properties under investigation is ursolic acid (UA; 3β-hydroxy-12-ursen-28-ic acid; Fig. 1A), a pentacyclic triterpene that is mainly found but not limited to peels of fruits and some herbs (7)(8)(9). Several beneficial effects of UA and its derivatives have been demonstrated in the treatment of various diseases, including cancer (10,11), obesity (12), glucose metabolism, and diabetes (13,14). UA is also known for its hepatoprotective (15,16), neuroprotective, and antidepressant (17,18) effects. The antiinflammatory activity of UA was examined in edema (8,19), passive cutaneous, and systematic anaphylaxis (20) and has been mainly attributed to its ability to affect extracellular signal-regulated kinase (ERK), NF-κB signaling pathways (21,22), and the activity of protein tyrosine phosphatase (PTP)1B (23). However, there are no data on the molecular mechanisms of the inhibitory effect of UA on the earliest stages of FcεRIsignalosome activation.
There are two fundamental theories of UA action on cells, protein-centric and lipid-centric. The protein-centric theory proposes that UA interacts with specific proteins and, in this way, affects their properties. Several UA-binding proteins have been described, including PTB1B (23), BCL2 (24), and αglucosidase (25). The lipid-centric theory postulates that UA, similarly to, for example, anesthetics (26), dissolves in and interacts with cellular lipids and acts by changing the physical properties and spatial arrangement of plasma membrane components. This theory is supported by X-ray diffraction, differential scanning calorimetry, 31 P NMR, Raman spectroscopy, and fluorescence anisotropy using liposomes of defined lipid composition (27)(28)(29)(30)(31).
In this study, we used mouse bone marrow-derived mast cells (BMMCs) and rat basophilic leukemia (RBL) cells and examined the sensitivity of FcεRI-initiated signaling events in the presence or absence of UA. We focused on tyrosine phosphorylation of the FcεRI-β subunit and other kinase and phosphatase substrates, calcium response, degranulation, and production and secretion of tumor necrosis factor (TNF-α). We also examined UA-induced changes in chemotaxis toward various chemoattractants and changes in the physical properties of the plasma membrane. Our data indicate that UA interferes with the earliest signaling events in FcεRI triggering. This effect seems to be, at least in part, the result of UAinduced changes in the properties of the plasma membrane. In this way, our data support the lipid-centric hypothesis of UA action at the earliest stages of mast cell signaling.

Short-term exposure to UA inhibits FcεRI-induced degranulation and calcium response
To examine the effects of short-term UA on antigeninduced activation of mast cells, we used BMMCs from C57BL/6 mice cultured for 8 to 12 weeks in culture media supplemented with stem cell factor (SCF) and interleukin (IL)-3. More than 95% of the cells were mast cells, as deduced from the expression of both FcεRI and KIT receptors (not shown). Exposure of the BMMCs for 15 min to UA at a concentration up to 75 μM had no toxic effect, as determined by the trypan blue dye exclusion test (not shown). Detailed analysis by flow cytometry (Fig. 1, B and C) showed that a small fraction of BMMCs exhibited significantly increased propidium iodide (PI) fluorescence (Q1) 15 min after exposure to 50 μM UA (7.2% ± 1.2%; mean ± SEM) or 75 μM UA (7.4% ± 1.1%). A small fraction of BMMCs also showed increased cell surface expression of phosphatidylserine, as reflected by an increased number of annexin V-FITC positive cells (Q3) 15 min after exposure to 50 μM UA (2.6% ± 0.4%) or 75 μM UA (3.1% ± 0.3%). Interestingly, only a small number of cells simultaneously exhibited both markers (Q2) after the treatment with 50 μM UA (2.5% ± 1.0%) or 75 μM UA (3.2% ± 0.8%). These and other data (see later) suggested that short-term pretreatment of the cells with UA at the concentrations used had minimal toxic effects. The observed increased staining with PI and annexin V-FITC could reflect changes in the plasma membrane caused by an interaction of UA with the plasma membrane lipids and not because of a toxic effect of the drug (32,33).
Pretreatment of IgE-sensitized BMMCs for 15 min with UA at 25 to 75 μM significantly inhibited antigen-induced degranulation in a dose-dependent way ( Fig. 2A). Almost complete inhibition of degranulation was observed with 75 μM UA. It should be noted that at all concentrations tested, UA did not affect the spontaneous release of β-glucuronidase from nonactivated cells. This observation supports the notion that short-term exposure of BMMC to UA at concentrations up to Figure 1. Increased phosphatidylserine plasma membrane expression and propidium iodide (PI) staining in BMMCs exposed to UA. A, chemical structure of UA. B, flow cytometry analysis of cells exposed for 15 min to BSS-0.1% BSA supplemented with 0.1% DMSO, used as a control (Ctrl), or UA at a concentration of 50 μM or 75 μM. The cells were stained for phosphatidylserine expression with annexin V-FITC conjugate and for membrane integrity with propidium iodide. The percentage of positive cells in individual quadrants (Q) from a typical experiment is also shown. C, data from three independent experiments performed as in (B) were analyzed, and mean values ± SEMs in quadrants Q1, Q2, and Q3 were calculated. Statistical significance of the intergroup differences is also shown. BMMC, bone marrow-derived mast cells; DMSO, dimethyl sulfoxide; UA, ursolic acid.
75 μM is not toxic to BMMCs. In further experiments, we used UA at a concentration of 50 μM, which inhibited degranulation by approximately 50%.
Mast cells could also be activated by thapsigargin, which induces the release of Ca 2+ from intracellular stores by inhibiting the endoplasmic reticulum Ca 2+ ATPase and, in this way, bypasses all FcεRI-mediated early activation steps. When BMMCs were exposed to 1 μM thapsigargin, strong degranulation was observed (Fig. 2B). Pretreatment with 50 μM UA for 15 min significantly inhibited this degranulation (Fig. 2B).
To determine whether the calcium response is also affected by 15 min exposure to UA, we monitored concentrations of free cytoplasmic calcium [Ca 2+ ] i with a Fura-2 probe. We found that 50 μM UA significantly inhibited the antigeninduced calcium response in cells kept in the presence of 1.8 mM calcium (Fig. 2C). Cells activated by antigen in the absence of extracellular calcium exhibited only a small increase in [Ca 2+ ] i , which was significantly inhibited by UA. After the addition of 1 mM calcium, a rapid rise in [Ca 2+ ] i was observed in control cells, as described in previous studies (34,35) but not in UA-treated cells (Fig. 2D). The reduced calcium response in UA-treated cells was, at least in part, caused by reduced uptake of extracellular Ca 2+ , as determined by 45 Cabinding assay. In this assay, antigen-induced activation led to an increase in 45 Ca uptake, which was significantly inhibited by 50 μM UA (Fig. 2E). When compared to antigen-activated cells, thapsigargin-activated cells exhibited higher 45 Ca uptake, which was inhibited by 50 μM UA to the levels comparable to those in UA-treated and antigen-activated cells (Fig. 2E).
Short-term exposure to UA inhibits chemotaxis toward antigen but not towards PGE 2 and SCF An essential aspect of mast cell physiology is chemotaxis toward various chemoattractants (36). Next, we, therefore, compared chemotaxis of vehicle-and UA-treated cells toward antigen, prostaglandin (PG)E 2 , and SCF. In the transwell migration assay, 15 min pretreatment of IgE-sensitized BMMCs with 50 μM UA resulted in significantly reduced chemotaxis toward antigen (Fig. 2F). When PGE 2 was used as a chemoattractant, chemotaxis was higher but was not significantly inhibited by 50 μM UA. No significant inhibition by UA was also observed when SCF was used as a chemoattractant (Fig. 2F). These data indicate that UA selectively blocks chemotaxis toward antigen and not the general ability of the cells to migrate toward chemoattractants.
Short-term exposure to UA promotes the production of TNF-α but inhibits its release from antigen-activated cells Antigen-mediated activation of BMMCs leads to enhanced secretion of cytokine gene products (1,2). To determine whether UA has any effect on this process, we analyzed the levels of TNF-α mRNA in IgE-sensitized BMMCs treated with 50 μM UA for 15 min and then activated with antigen in the presence of UA for 60 min. We found that treatment with UA in nonactivated cells did not affect basal levels of TNF-α mRNA (Fig. 3A). Surprisingly, in antigen-activated cells, the amount of TNF-α mRNA in UA-treated cells was significantly , which was also present during antigen-mediated activation. Fifteen minutes after adding antigen (Ag; 250 ng/ml TNP-BSA in BSS-0.1% BSA) or BSS-0.1% BSA alone (Ctrl), degranulation was quantified by analysis of β-glucuronidase levels. B, BMMCs were exposed for 15 min to 50 μM UA or vehicle (0.1% DMSO). The cells were activated with 1 μM thapsigargin (Th) or were exposed to BSS-0.1% BSA (Ctrl). Degranulation was determined after 15 min as above. C, IgE-sensitized and Fura-2-loaded BMMCs were preincubated with 50 μM UA or vehicle (0.1% DMSO; Ctrl) in BSS containing 1.8 mM Ca 2+ . After 15 min, the cells were activated by antigen (Ag; 250 ng/ ml TNP-BSA) at the time point indicated by an arrow. [Ca 2+ ] i was monitored for the indicated time intervals as changes in fluorescence ratios of 340/ 380 nm. D, the cells were treated the same way as in (C), except that BSS without calcium was used, and 1.8 mM calcium was added at the indicated time interval after antigen triggering (arrow, Ca 2+ ). E, IgE-sensitized cells were treated for 15 min with vehicle (Ctrl) or 50 μM UA and then activated with antigen (250 ng/ml TNP-BSA) or thapsigargin (1 μM) in the presence of extracellular 45 Ca 2+ (1 mM). After 15 min at 37 C, the reaction was terminated by centrifugation of the cells through 12% BSA in PBS and cell-bound radioactivity was determined. F, the effect of UA on BMMC chemotaxis was determined in transwell chambers. IgE-sensitized cells were treated for 15 min with vehicle (Ctrl) or 50 μM UA and then transferred into upper wells of transfer chambers. Migration of the cells toward antigen (250 ng/ml), PGE 2 (100 nM), or SCF (100 ng/ml), present together with 50 μM UA in bottom wells, was determined after 6 h. Means ± SEMs were calculated from 3 (C and D) to 4 to 17 (A, B, E, and F) independent experiments. Statistical significance of intergroup differences is also shown. BMMC, bone marrow-derived mast cells; BSA, bovine serum albumin; BSS, buffered salt solution; DMSO, dimethyl sulfoxide; TNP, 2,4,6-trinitrophenol; UA, ursolic acid.
higher compared to the cells treated with vehicle (0.1% dimethyl sulfoxide [DMSO]).
In UA-treated and antigen-activated cells, the amount of TNF-α released to the supernatant was significantly reduced when compared to antigen-activated control cells (Fig. 3B). To determine whether the reduced amount of secreted TNF-α in UA-treated cells reflects the decreased amount of total cellular TNF-α, we examined the presence of TNF-α by flow cytometry in saponin-permeabilized cells. We found a significantly increased percentage of TNF-α-positive cells in the UAtreated and antigen-activated cells than in activated control cells (Fig. 3, C and D). In antigen-activated and vehicle-treated cells, the geometric mean of fluorescence intensity (geo MFI) was 1143 ± 70 (mean ± SEM). In cells treated with UA and activated with antigen, the geo MFI was increased 3.7 times to 4250 ± 506; the difference was significant (p < 0.002). When the assay was performed with cells not permeabilized with saponin to measure surface TNF-α, we found only 1.4 higher geo MFI of TNF-α at the cell surfaces of activated and UAtreated cells when compared to activated control cells, and the difference was not significant (Fig. 3E). The data suggested that the reduced secretion of TNF-α is not caused by an accumulation of pro-TNF-α at the plasma membrane due to reduced activity of ADAM metallopeptidase, an enzyme responsible for pro-TNF-α cleavage and mature TNF release (37).
Different effects of UA on tyrosine phosphorylation of various signal transduction proteins in FcεRI-activated cells The first biochemically well-defined step in FcεRI signaling is tyrosine phosphorylation of FcεRI β and γ subunits by the LYN kinase, followed by the formation of the FcεRI signalosome and enhanced phosphorylation of numerous substrates (38,39). Next, we therefore examined tyrosine phosphorylation of proteins in nonactivated or antigen-activated cells pretreated with UA or vehicle. First, we analyzed global tyrosine-phosphorylated proteins in cell lysates sizefractionated by SDS-PAGE, followed by immunoblotting with a phosphotyrosine-specific antibody PY-20-horseradish peroxidase (HRP) conjugate. As expected, in the absence of UA, antigen activation resulted in increased phosphorylation of several proteins on tyrosines (Fig. 4A). Treatment of the cells with UA alone (without antigen activation) also induced increased tyrosine phosphorylation of a similar set of proteins. FcεRI-mediated activation of UA-treated cells did not cause a further increase in global tyrosine phosphorylation.
Next, we examined phosphorylation of individual proteins using SDS-PAGE-fractionated whole cell lysates followed by immunoblotting with phosphoprotein-specific antibodies. Alternatively, we immunoprecipitated proteins of interest and analyzed them by immunoblotting with the mAb PY-20-HRP conjugate. To determine whether UA interferes with tyrosine phosphorylation of the FcεRI β subunit, we immunoprecipitated FcεRI from UA-or vehicle-treated cells that were activated or not with antigen. UA had no significant effect on tyrosine phosphorylation of the FcεRI β in nonactivated cells (Fig. 4, B and C). In antigen-activated cells, tyrosine phosphorylation of FcεRI-β was significantly higher in UApretreated cells than in control cells (Fig. 4, B and C).
An essential protein-tyrosine kinase in the FcεRI signaling is SYK, which is phosphorylated on several residues in antigenactivated cells. Tyr 519 and Tyr 520 of mouse SYK (equivalent to Tyr 525/526 of human Syk) are located in the activation loop of the kinase, and their phosphorylation is crucial for SYK functioning (40). We found that 15 min treatment of BMMCs with UA induced significantly enhanced phosphorylation of Tyr 519/520 in both nonactivated and antigen-activated cells in all time intervals analyzed (Fig. 4, D and E). SYK kinase phosphorylates the linker for activation of T cells (LAT)1 and Plasma membrane as an ursolic acid target LAT2 adapter proteins in antigen-activated cells (41); therefore, it was not surprising to find enhanced tyrosine phosphorylation of LAT1 and LAT2 in antigen-nonactivated and UA-treated cells when compared to vehicle-treated cells. After stimulation with antigen in vehicle-treated cells, phosphorylation of LAT1 and LAT2 was increased as expected. Surprisingly, there was no further increase in LAT1 and LAT2 tyrosine phosphorylation after antigen stimulation in UAtreated cells (Fig. 4, F-I).
It is known that phosphorylation of PLCγ at Tyr 783 by SYK kinase enhances the enzymatic activity of PLCγ (42). In agreement with enhanced phosphorylation of SYK kinase, phosphorylation of PLCγ Tyr783 was increased in control and antigen-activated cells treated with UA. Enhanced phosphorylation of PLCγ Tyr783 was highly significant in all time intervals (Fig. 5, A and B).
A critical signal transduction molecule involved in the activation of many transcription factors is ERK. To determine whether UA is involved in the phosphorylation of ERK Tyr204 , which correlates with ERK activation, we examined ERK phosphorylation in control and UA-treated cells. We found significantly increased phosphorylation of ERK Tyr204 in UA-treated nonactivated cells when compared to control cells. After activation of the cells by FcεRI triggering, ERK Tyr204 phosphorylation was significantly lower in UA-treated cells than UAuntreated cells in all time intervals analyzed (Fig. 5, C and D).
We also analyzed phosphorylation of E3 ubiquitin-protein ligase, CBL, and mammalian target of rapamycin (mTOR) proteins. The tyrosine-phosphorylated CBL protein is involved in the negative regulation of FcεRI-mediated signaling by promoting ubiquitination of the activated receptor subunits and associated protein tyrosine kinases (43,44). Pretreatment of BMMCs with 50 μM UA reduced phosphorylation of CBL Tyr700 in both nonactivated cells and cells activated with antigen for 1 min. Cells activated for 5 or 15 min and treated or not with UA showed similar phosphorylation of CBL Tyr700 (Fig. 5, E and F). mTOR is a serine/threonine-protein kinase, which is phosphorylated on Ser 2448 via the PI3 kinase/Akt signaling pathway and regulates mast cell degranulation, cytokine production, and migration (45)(46)(47). We found significantly higher levels of phosphorylated mTOR S2448 in nonactivated UA-treated cells than in control nonactivated cells. After stimulation with antigen, tyrosine phosphorylation was lower in UA-treated cells than in cells treated with vehicle ( Fig. 5, G and H).
Finally, we examined the phosphorylation of the LYN kinase. The catalytic activity of this enzyme is regulated in mice by phosphorylation/dephosphorylation of two conserved tyrosines at positions 397 and 508 (48)(49)(50). Phosphorylation at Tyr 397 stabilizes the activation loop of the catalytic domain and increases LYN activity. In contrast, phosphorylation of Tyr 508 Plasma membrane as an ursolic acid target at the C terminus of LYN by C-terminal Src kinase (CSK) stabilizes the inactive conformation of the catalytic domain. We found similar levels of LYN with phosphorylated Tyr 397 and Tyr 508 in UA-treated and control cells and no significant changes in phosphorylation of these regulatory tyrosines in antigen-activated cells (Fig. S1, A-C). The data suggest that the observed changes in phosphorylation of the FcεRI β subunit (Fig. 4, B and C) is not due to the enhanced activity of the LYN kinase. Rather, UA-induced changes in the spatial arrangement of plasma membrane molecules and their mobility could be involved.
Short-term exposure to UA induces changes in the properties of detergent-resistant membrane domains Previous studies with mast cells showed that acute cholesterol lowering or changes in the plasma membrane lipid composition resulted in changes in the properties of detergentresistant membranes (DRMs) (35,(51)(52)(53). In further experiments, we examined whether UA induces changes in the distribution of tyrosine-phosphorylated proteins in DRMs. There is a relatively small fraction of cellular proteins, such as transmembrane adapter proteins (phosphoprotein associated with glycosphingolipid-enriched microdomains (PAG), LAT1, and LAT2) and LYN kinase, which are localized in DRMs (41,54). In nonactivated BMMCs, these proteins exhibit tyrosine phosphorylation at relatively low levels (LAT1, LAT2) or high levels (PAG, LYN; Fig. S2, A and E). Exposure of the cells to UA resulted in an increased fraction of tyrosinephosphorylated LYN, LAT1, and LAT2 in DRMs (Fig. S2, B and E). Activation with antigen in control cells reduced the fraction of phosphorylated LYN and PAG in DRMs (Fig. S2, C and E), as expected (55,56). In cells pretreated with UA and activated with antigen, we observed DRMs with significantly increased levels of phosphorylated LAT1 and LAT2 (Fig. S2, D and E). These data indicate that UA interferes with the distribution of tyrosine-phosphorylated proteins in DRMs.
A previous study identified UA as a cholesterol-lowering drug (57). To determine whether short-term treatment of the cells to UA interferes with surface cholesterol exposure, we used filipin as a probe (58) to quantify plasma membrane surface cholesterol. In the absence of filipin, BMMCs exhibited only weak autofluorescence (Fig. 6, A and B; Ctrl-F). In the presence of filipin, the cells showed fluorescence that was similar to the control cells (Ctrl+F), cells pretreated with vehicle (0.1% DMSO+F), or containing 50 μM UA and 0.1 % DMSO. As a control, we also included cells pretreated with methyl-β-cyclodextrin (MβCD), which removes cholesterol from the cells and disrupts detergent-insoluble glycolipid domains (51,59,60). We found that MβCD-treated cells exhibited significantly reduced filipin fluorescence.

Reduced patches formation of aggregated FcεRI and Thy-1 in UA-treated cells
In our previous study, we found that pretreatment of RBL-2H3 cells with MβCD, which removes plasma membrane cholesterol, resulted in a transient increase in tyrosine phosphorylation of several proteins, increased antigen-induced and spontaneous degranulation, and increased plasma membrane mobility of the antibody-aggregated glycosylphosphatidylinositol-anchored protein Thy-1.1 (52). In further experiments, we examined the effect of UA on the plasma membrane mobility of the antibody-aggregated FcεRI and Thy-1.1 glycoprotein in, respectively, BMMCs and RBL-2H3 cells. The data show that pretreatment of BMMCs with UA reduced the mobility and formation of patches of aggregated FcεRI (Fig. 7, A and B). Reduced mobility by UA was also observed when antibody-mediated aggregation of the Thy-1.1 glycoprotein was analyzed (Fig. 7, C and D).

Reduced mobility of FcεRI and cholesterol in UA-treated cells
To quantify the effect of UA on the lateral diffusion of the FcεRI and cholesterol, the fluorescence recovery after the photobleaching (FRAP) method was used. When RBL-2H3 cells were exposed for 15 min to 50 μM UA, relative fluorescence of the FcεRI-IgE-FITC complexes in the photobleached area was significantly reduced when compared to cells pretreated with DMSO (vehicle; Fig. 8, A and B). From the FRAP data obtained, we calculated that control cells pretreated with vehicle (DMSO) showed a lower number of immobile FcεRI-IgE-FITC complexes (34.40 ± 3.22 %; mean ± SEM) when compared to cell pretreated with UA (64.00 ± 5.16 %). The intergroup difference was significant (p < 0.001).

Discussion
Data presented in this study show that short-term exposure of BMMCs to UA interferes with antigen-mediated effector functions. To understand the molecular mechanisms of the effect of UA on mast cell activation events, we examined various signaling pathways and targets. Several lines of evidence suggest that UA interferes with the function of the FcεRI-signalosome and support the lipid-centric theory of UA action at least in the earliest stages of mast cell activation.
Pretreatment of mast cells with UA in the concentration range of 10 μM to 75 μM inhibited the release of β-glucuronidase after FcεRI triggering in a dose-dependent manner. In this dose range, no increase in the spontaneous release of βglucuronidase was observed, suggesting that UA had no toxic effect. These data are seemingly in contrast with the results of experiments showing that 15 min exposure of BMMCs to UA at 50 μM and even more at 75 μM, resulted in increased surface expression of phosphatidylserine as detected with annexin V-FITC and increased staining with PI. Phosphatidylserine is usually restricted to the inner leaflet of the plasma membrane, and this asymmetry is maintained by the activity of energy-dependent flippases and floppases, which mediate, respectively, inward-directed and outward-directed transfer of phospholipids (61)(62)(63). The equilibration of phospholipids between the two plasma membrane leaflets seems to be regulated by lipid scramblase, which facilitates the bidirectional migration of phospholipids across the lipid bilayers (64,65). Flippases, floppases, and scramblases are transmembrane proteins, and therefore, their activity could be affected by the insertion of UA into the plasma membrane. Enhanced staining with PI could also be related to UA-induced changes in the plasma membrane properties rather than cell death. Deitch et al. have shown that PI does not bind only to dsDNA in the cell nucleus but also to dsRNA (66), and this could lead to an increased number of false-positive events that are associated with PI staining of RNA within the cytoplasmic compartment (32). Thus, changes in the plasma membrane properties caused by incorporating hydrophobic UA into the plasma membrane could be responsible for increased staining of cells with PI.
Detailed analysis of protein tyrosine phosphorylation in nonactivated and antigen-activated cells showed that UA dramatically affected early activation events. Interestingly, several proteins, including SYK, LAT1, LAT2, PLCγ, ERK, and mTOR, showed increased tyrosine phosphorylation induced by 15 min treatment with UA alone. These proteins are rapidly activated after triggering the cells with antigen and belong to the so-called FcεRI signalosome (39,67). Interestingly, after stimulation with antigen, UA-treated cells, when compared to vehicle-treated cells, exhibited significantly enhanced or reduced tyrosine phosphorylation of some of these signaling proteins. One of the first well-defined enzymatic steps in antigen-activated mast cells is LYN kinase-mediated tyrosine phosphorylation of the FcεRI β subunit (68,69). UA alone did not affect phosphorylation of this substrate but significantly enhanced phosphorylation of the FcεRI β subunit after antigen-mediated triggering. The FcεRI β subunit is phosphorylated by LYN kinase, which showed no significant changes in phosphorylation of Tyr at positions 397 and 508, serving, respectively, as a positive and a negative regulator of the LYN kinase activity. The second well-defined step in FcεRI signaling is the binding of SYK to the FcεRI γ subunit and SYK activation by its phosphorylation, followed by phosphorylation of numerous SYK substrates (70,71). We found enhanced SYK phosphorylation in cells exposed to UA alone, and this phosphorylation was further increased in antigen-activated cells.
Enhanced phosphorylation of SYK in cells treated with UA alone was connected with increased phosphorylation of LAT1, LAT2, and PLCγ, important substrates of SYK. However, the higher phosphorylation of PLCγ was not accompanied by an increased calcium response. In fact, the levels of [Ca 2+ ] i were reduced in UA-pretreated and antigen-activated cells, which was likely responsible for impaired degranulation. The lower [Ca 2+ ] i in UA-treated and FcεRI-activated cells was mainly caused by reduced calcium uptake, as determined in experiments measuring 45 Ca uptake. These data were supported by reduced [Ca 2+ ] i after adding Ca 2+ to cells activated by antigen in Ca 2+ -free buffered salt solution (BSS). Similar inhibition of degranulation and calcium response was observed in UAtreated cells activated by thapsigargin. Since thapsigargin bypasses all early FcεRI-mediated signaling steps, it is likely that a key factor leading to UA-mediated inhibition of degranulation in antigen-activated cells is due to reduced Ca 2+ uptake. The negative effect of UA on calcium channel activity could be related to UA-mediated changes in lipid bilayer properties (72). In fact, previous studies showed that compounds that increase rigidity (73) or reduce fluidity (74) of the plasma membrane inhibit store-operated Ca 2+ (SOC) channels. Insertion of UA into the plasma membrane could inhibit SOC channels through direct interaction of UA with Orai1 protein in a similar way as was described for inhibition of Orai1 by cholesterol (75).
It should be noted that UA negatively regulated only some of the activation events. Thus, UA-treated cells exhibited reduced chemotaxis toward antigen but not toward PGE 2 or SCF. Although the molecular basis of these differences is not known, it could be, at least in part, dependent on UA-mediated changes in the plasma membrane properties. UA is a multicyclic rigid planar structure with a hydrophilic 3β-hydroxyl group. In its structure, UA is similar to cholesterol, which is also a multicyclic rigid planar structure with a hydrophilic 3βhydroxyl group and a hydrophobic side chain. Thus, both UA Figure 7. UA interferes with the patches formation after FcεRI and Thy-1.1. crosslinking. A and B, formation of FcεRI patches. A, IgE-sensitized BMMCs were washed and exposed for 15 min to 50 μM UA or vehicle (Ctrl). Then, the cells were exposed to AF 488-conjugated secondary antibody for 10 min at 37 C in the presence of 50 μM UA or vehicle (Ctrl), followed by fixing with 4% paraformaldehyde. Alternatively, the cells were fixed before exposure to AF 488conjugated secondary antibody for 10 min. The fluorescence of cells was examined by confocal microscopy. B, the distribution of FcεRI in individual cells was evaluated, and the fraction of fluorescence detected centrally was calculated. Means ± SEMs and statistical significance of the intergroup differences are presented. C and D, formation of Thy-1.1 aggregates. C, adherent RBL-2H3 cells were treated with anti-Thy-1.1 antibody. After 30 min, the antibody was washed out, and the cells were exposed to 50 μM UA or vehicle as above. Then, the cells were fixed before (0 min) or after (10 min A and B, mobility of the FcεRI-IgE-FITC complexes. A, RBL-2H3 cells were exposed for 15 min to 50 μM UA in DMSO (UA) or the corresponding concentration of vehicle (Ctrl). Then the cells were exposed to IgE-FITC complexes. Image sequences are from cells at the indicated time points before photobleaching (Pre-bleach) at photobleaching (Bleach) or after photobleaching (Post-bleach; the circled areas). The graph in (B) quantifies the recovery into the photobleached areas. Means ± SEMs were calculated from 18 cells in two independent experiments with similar results. C and D, mobility of the cholesterol-Filipin complexes. C, RBL-2H3 cells were exposed for 15 min to 50 μM UA in DMSO or the corresponding concentration of vehicle (Ctrl). Then the cells were exposed to filipin at 50 μg/ml for 10 min. After washing, the image sequences from cells at the indicated time points before, at, or after photobleaching of the circled areas. The graph in (D) quantifies the recovery into the photobleached region. Means ± SEMs were calculated from 15 cells in two independent experiments with similar results. Statistical significance of differences between control (Ctrl) and UA-treated cells (UA) is also indicated. Bar = 10 μm. DMSO, dimethyl sulfoxide; UA, ursolic acid. and cholesterol can tightly assemble the molecules of the phospholipid bilayer and cause the membrane to be less fluid (76,77). Furthermore, regarding its cell membrane-associated functions, cholesterol is also implicated in the modulation of cellular signal transmission and intracellular trafficking by contributing to lipid raft assemblies (78)(79)(80). Comparable properties of cholesterol and UA in their fluidity-modulating and condensing effects were also shown in studies using liposomal membranes (27). In these studies, UA was shown to locate near the surface in the vicinity of the phospholipid headgroups, with orientation depending on the lipid composition. In complex lipid systems, UA interacted specifically with POPE (1-palmitoyl-2-oleoyl-sn-glycero-3phosphoethanolamine), PSM (N-palmitoyl-D-erythro-sphingosyl-phosphorylcholine), and cholesterol and excluded POPG (1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1 0 -rac-glycerol) from its surrounding (28,31). Furthermore, UA reacted specifically with phosphatidylethanolamine phospholipids through hydrogen bonding, disorganizing the membrane interface and thus disrupting the membrane integrity (81). The combined data suggest that UA, similarly to other triterpenoids and hydrophobic bioactive molecules, incorporates into biological membranes and modulates their fluidity, morphology, and permeability (29,82). Based on the similarity of cholesterol and UA, it is not surprising that chemotaxis of mast cells toward antigen and PGE 2 showed opposite effects when the cells were pretreated with UA or MβCD (83,84). The similarity between UA and cholesterol could also explain the increased levels of TNF-α in antigen-activated mast cells pretreated with UA (this study) or cells activated by cholesterol with ionomycin (85,86). However, as shown in our study, the release of TNF-α from UA-treated cells is reduced. This could be caused by reduced [Ca 2+ ] i , which is important for TNF-α exocytosis (87).
The activation processes are initiated in DRM microdomains called lipid rafts. In nonactivated cells, pretreatment with UA for 15 min resulted in an increased amount of tyrosine-phosphorylated LYN, LAT1, and LAT2 in DRMs. In antigen-activated cells, pretreatment with UA resulted in an enhanced amount of tyrosine-phosphorylated LYN and LAT2 in DRMs. These data suggest that the properties of lipid microdomains are disorganized by UA, which could have contributed to the observed changes in the signaling properties of the FcεRI signalosome.
In experiments in which we followed the formation of patches of aggregated FcεRI and Thy-1.1 glycoprotein, we obtained evidence that UA reduced aggregation of the surface components. In connection with these experiments, it is relevant that lowering plasma membrane cholesterol with MβCD resulted in acceleration of Thy-1 aggregation and rapid formation of caps (88). Direct evidence that mobility of plasma membrane components is reduced by UA was obtained in FRAP experiments in which UA significantly increased the immobile fraction of FcεRI and cholesterol.
In most of the experiments in this study, we used UA at a concentration of 50 μM. Although this is a relatively high concentration, it can be clinically relevant when the drug is administered topically. In the mouse model, oral administration of UA suppressed both IgE-mediated passive cutaneous anaphylaxis and ovalbumin-induced active systemic anaphylaxis at concentrations similar to those of dexamethasone, which is clinically used (20).

Conclusions
As depicted in Figure 9, there is a coordinated synergy of lipid-based and protein-based interactions in the earliest stages of FcεRI signaling. In nonactivated cells, a large fraction of LYN kinase preferentially partitions into liquid-order (Lo)-like lipidic nanodomains in the plasma membrane due to its saturated lipid anchor (palmitoyl and myristoyl). Lo-like domains also carry other components of the FcεRI signalosome, such as transmembrane adapter proteins LAT1, LAT2, and PAG. IgE receptors seem to be separated into liquiddisordered (Ld)-like nanodomains, where phosphatases keep phosphorylation of the FcεRI subunits at nonfunctional thresholds (39,41,89). Activation of IgE-sensitized cells by multivalent antigen leads to aggregation of antigen-IgE-FcεRI complexes, the enhanced association of the FcεRI aggregates with Lo-like domains, and reduced physical and functional association of the phosphatases with the FcεRI complexes. These changes alter the balance between kinases and phosphatases in favor of kinases, leading to increased phosphorylation of the FcεRI subunits (39). The phosphorylated FcεRI γ subunit is a transient anchor of SYK kinase that phosphorylates adapter proteins and contributes to further propagation of the activation signal, resulting in enhanced calcium response, degranulation, and cytokine production. UA is a hydrophobic pentacyclic triterpenoid, which incorporates into the plasma membrane and changes its properties, including mobility of the plasma membrane components. In this way, UA itself could change the balance between kinases and phosphatases, leading to tyrosine phosphorylation of numerous substrates without receptor triggering. Activation of UA-pretreated cells with antigen leads to further phosphorylation of some kinase substrates (FcεRI β subunit, SYK kinase, PLCγ) or reduced phosphorylation of other signaling proteins (LAT1, ERK, CBL, mTOR). UA-dysregulated signaling events could lead to a reduced calcium response and degranulation or increased production of cytokines. Although UA could bind to and modify some signaling proteins, the earliest FcεRI-activation events affected by UA seem to reflect a direct interaction of UA with the plasma membrane lipids. The combined data suggest that UA induces plasma membrane perturbations by directly interacting with the plasma membrane lipids and support the lipid-centric hypothesis of UA action at the initial stages of mast cell activation.

Mice and cells
Mice were bred and maintained in a specific pathogen-free facility of the Institute of Molecular Genetics, used in compliance with the Institute guidelines, and approved by the Institute Commission for Animal Welfare and Protection (Approval No 23/2013). BMMC precursors were isolated from the femurs and tibias of 6-to 8-week-old mice of the C57BL/6 genotype. They were cultured for 8 to 12 weeks in a complete growth medium consisting of RPMI-1640 medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 71 μM 2-mercaptoethanol, minimal essential medium, nonessential amino acids, 0.7 mM sodium pyruvate, 2.5 mM Lglutamine, 12 mM D-glucose, recombinant mouse SCF (15 ng/ ml), recombinant mouse IL-3 (15 ng/ml), and 10% (vol/vol) fetal calf serum. In experiments presented in Figs. 3 and 7, we used cells of the BMMC line (BMMCL) (97), which were cultured under the same conditions as BMMCs, except that SCF was omitted from the culture medium. For experiments presented in Figs. 7 and 8, we also used RBL cells, clone 2H3 (98), which were cultured as previously described (99).

Annexin V-FITC and PI staining
The annexin V-binding assay was used for quantification of phosphatidylserine surface expression in BMMCs treated with various concentrations of UA. Briefly, the cells were washed with BSS (20 mM Hepes, pH 7.4, 135 mM NaCl, 5 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5.6 mM glucose) supplemented with 0.1% BSA. The assay was performed on a 96-well plate. Cells (0.25 × 10 6 ) in 100 μl BSS-0.1% BSA were exposed to various concentrations of UA for 15 min at 37 C, centrifuged at 4 C and washed with ice-cold 1× annexin-binding buffer. The cell pellets were resuspended in 50 μl of ice-cold 1× annexin V-binding buffer supplemented with 5 μl annexin V-FITC or 1× annexin V buffer alone for unstained control cells. Cells were incubated for 30 min on ice in the dark, and 5 min before flow cytometry analysis, 50 μl of the binding buffer with or without PI was added. Samples were measured in LSRII FACS DIVA (BD Bioscience). Annexin V-FITC binding was analyzed by flow cytometry using the FITC signal detector (FL1) and PI staining by the phycoerythrin emission detector (FL2).

Cell activation, degranulation, and calcium response
BMMCs were sensitized in the complete growth medium without IL-3 and SCF, supplemented with IgE (1000× diluted ascites formed by hybridoma cells IGEL b4.1). After 12 to 14 h of incubation, cells were washed and resuspended in BSS-0.1% BSA at a concentration of 1.5 × 10 6 cells/ml. Cells (100 μl aliquots) were transferred into wells of a 96-well cell culture plate. The cells were pretreated for 15 min with various concentrations of UA or vehicle (0.1% DMSO) in BSS-0.1% BSA added in 100 μl aliquots and challenged with 20 μl of antigen (TNP-BSA conjugate) at a concentration of 250 ng/ml. When the UA-treated cells were activated by thapsigargin, the sensitization step was omitted. The degree of degranulation was determined as the amount of β-glucuronidase released into the supernatants as described (56) using 4methylumbelliferyl-β-D-glucuronide hydrate as a substrate. For the total amount of β-glucuronidase, the cells were treated with 0.5% Triton X-100. Fluorescence was determined by an Infinite M200 microtiter plate reader (Tecan) at 355-nm excitation and 460-nm emission filters.
Changes in the [Ca 2+ ] i were analyzed using cytoplasmic reporter Fura-2-AM as described (56). Briefly, IgE-sensitized cells were incubated with Fura-2-AM in BSS-0.1% BSA supplemented with 2.5 mM probenecid (35). After 30 min, cells were washed and treated with UA or vehicle in the presence of probenecid for 15 min. Before the measurement, the cells were washed and resuspended in BSS-0.1% BSA without probenecid. The free cytoplasmic Ca 2+ levels were determined by an Infinite M200 microtiter plate reader in a NUNC white 96-well plate (Thermo Scientific) with excitation wavelengths at 340 and 380 nm and constant emission at 510 nm. To determine the effect of UA on antigen-induced changes in [Ca 2+ ] i in the absence of extracellular calcium, the same method was used, except that BSS without CaCl 2 was used, and after antigen triggering, CaCl 2 was added as indicated in Figure 2D.
Extracellular calcium uptake was measured as previously described (53). Briefly, IgE-sensitized or nonsensitized cells (2 × 10 6 ) were washed in BSS-0.1% BSA, resuspended in 100 μl BSS-0.1% BSA with 1 mM Ca 2+ , and mixed with 100 μl UA. After 15 min incubation at 37 C, the cells were activated for 10 min at 37 C by adding 20 μl BSS-0.1% BSA supplemented with a mix of 45 Ca 2+ and antigen (250 ng/ml; IgE-sensitized cells) or thapsigargin (1 μM; IgE nonsensitized cells). The reaction was terminated by placing the tubes on ice, and cells with bound 45 Ca 2+ were separated from free 45 Ca 2+ by centrifugation at 1200g for 15 min at 4 C through 12% BSA as described (100). Cell pellets were recovered by freezing the tubes, slicing off the tube bottom, and solubilizing the cells with 1 ml of 1% Triton X-100. The radioactivity of the samples was measured in 10 ml scintillation liquid (EcoLite; ICN Biomedicals) using a scintillation counter with QuantaSmart software (PerkinElmer).

Cell lysates, immunoprecipitation, and immunoblotting
IgE-sensitized cells were treated or not with UA for 15 min at 37 C, followed by activation of the cells with antigen (TNP-BSA, 250 ng/ml) for different time intervals. Toward the end of the activation period, the cells were centrifuged and the pellet was resuspended in ice-cold immunoprecipitation buffer (25 mM Tris-HCl pH 8.0, 140 mM NaCl, 1 mM Na 3 VO 4 , 2 mM EDTA, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 mM PMSF) supplemented with 0.2% Triton X-100 (for FcεRI immunoprecipitation). Lysis buffer of the same composition but with 25 mM Tris-HCl, pH 7.5, supplemented with 1% n-dodecyl-β-D -maltoside and 1% Nonidet P-40 was used for LAT1 and LAT2 immunoprecipitations. After incubation on ice for 30 min, the lysates were centrifuged (16,000g for 5 min at 4 C), and the proteins in the postnuclear supernatants were immunoprecipitated with the corresponding antibodies prebound to UltraLink-immobilized protein A (Pierce, Thermo Scientific). The immunoprecipitated proteins were size fractionated by SDS-PAGE and analyzed by immunoblotting with PY-20-HRP conjugate or with protein-specific antibodies, followed by HRP-conjugated antimouse or anti-rabbit IgG antibodies. Several tyrosine-phosphorylated proteins were quantified by direct immunoblotting of SDS-PAGE-fractionated whole cell lysates with phosphoprotein-specific antibodies. For loading controls, protein-specific antibodies were used. The primary antibodies were detected with HRP-conjugated secondary antibodies. The HRP signal was detected by Luminescent Image Analyzer LAS-3000 (Fujifilm) and quantified with AIDA software (Raytest GmbH). The levels of phosphorylated proteins were normalized to the loading controls run in parallel gels (56).

Sucrose density gradient fractionation
Sucrose density gradient separations were performed as previously described (101). Briefly, IgE-sensitized BMMCs (15 × 10 6 ) were pretreated or not with UA for 15 min and then activated or not with antigen (250 ng/ml) for 5 min. After centrifugation, the cells were lysed on ice in 0.8 ml lysis buffer (10 mM Tris-HCl, pH 8.0, 50 mM NaCl, 2 mM EDTA, 10 mM glycerophosphate, 1 mM Na 3 VO 4 , 1 mM PMSF, 0.5 U/ml aprotinin, and 0.5 U/ml leupeptin) supplemented with 1% Brij-96. The gradient was formed by adding 0.5 ml of 80% sucrose stock solution to the bottom of a polyallomer tube (13 × 51 mm; Beckman Instruments) followed by 1.5 ml of 40% sucrose containing the cell lysate, 2 ml 30%, and 1 ml 5% sucrose. After 4 h centrifugation at 4 C (210,000g using an SW55 Ti rotor), 10 fractions were collected from the top of the gradient, and individual fractions were analyzed by SDS-PAGE followed by immunoblotting with PY-20-HRP and proteinspecific antibody.

Quantification of cell surface cholesterol
After 30 min incubation at 37 C with shaking, the cells were washed and fixed for 10 min at 22 C (in the dark) with 4% paraformaldehyde in staining buffer. The fixed cells were washed twice in staining buffer, exposed or not to 50 μl filipin (50 μg/ml PBS), and incubated for 30 min in the dark at 22 C.
After incubation, filipin was washed out twice, the cells were resuspended in staining buffer, and cholesterol levels were quantified using a BD FACSymphony flow cytometer with UV light laser (355 nm) and UV detector (450/50 nm).

Detection of cytokine mRNAs
IgE-sensitized or nonsensitized cells of the BMMC line were pretreated with 50 μM UA or vehicle (0.1% DMSO) for 15 min and then activated or not with antigen (TNP-BSA; 250 ng/ml) for 60 min. For cytokine mRNA detection, we used the RNAmediated oligonucleotide annealing, selection, and ligation (RASL)-quantitative PCR (qPCR) protocol, adapted from the miRNA detection method described previously (102). Briefly, RNA was isolated using a MicroElute Total RNA Kit (Omega Biotek) according to the manufacturer's instructions. Two DNA oligonucleotide probes (A and B, see later) specific for each cytokine mRNA were designed and synthesized. The oligonucleotide probes contained 15 to 16 nucleotide sequence regions complementary to a particular mRNA and a sequence for qPCR primer annealing (19 nucleotides). One of the probes (oligonucleotide B) has a phosphate on its 5 0 end to enable its ligation with the second probe (oligonucleotide A). The complementary sequences of both oligonucleotides are designed in such a way that they hybridize to mRNA right next to each other and the whole sequence is intron spanning. Both pairs of oligonucleotides (10 fmol each) for the given mRNA were mixed with the RNA sample (50 ng) and hybridized in 100 mM Tris-HCl pH 7.5 and 20 mM MgCl 2 for 2 min at 85 C, followed by 1 min incubation steps at temperatures 5 C lower than in the preceding step with the final step at 25 C. Next, 10 μl of the hybridization reaction was mixed with 1.25 U of SplintR ligase (New England BioLabs), 0.14 μl of 100 mM ATP, and 4 μl of 2× concentrated TP buffer (Top-Bio) containing 15 mM Tris-HCl, pH 8. Bead-based immunoassay for quantification of TNF-α released into supernatants BMMCs (2 × 10 6 ) were sensitized with IgE, then exposed to 50 μM UA or vehicle for 15 min, followed by activation or not with 250 ng/ml of TNP-BSA for 4 h. Then, the cells were centrifuged, and the supernatants were used for TNF-α quantification using the bead-based immunoassay (LEGENDplex multianalyte flow assay; mouse B effector panel 8-plex [BioLegend; catalog no.: 740820]) according to manufacturer's instructions. BD Symphony flow cytometer equipped with 488 and 637 nm lasers was used for data collection. The data were processed with LEGENDplex data analysis software.

Quantification of cellular and surface TNF-α
To quantify cellular TNF-α, the IgE-sensitized BMMCs were pretreated for 15 min with 50 μM UA or vehicle and then activated with antigen (TNP-BSA). After 2.5 h, the cells were centrifuged at 805g for 3 min at 21 C, followed by fixing with paraformaldehyde (4% in PBS) for 15 min at 21 C. After washing in PBS, the cells were permeabilized with 0.1% saponin and further processed as described (105). To quantify TNF-α exposed on the plasma membrane, we omitted the permeabilization of the cells with saponin.
FcεRI and Thy-1.1 mobility assay For the FcεRI mobility assay, 15-well multitest microscopy slides (MP Biomedicals) were coated overnight with fibronectin (10 μg/ml in PBS), blocked with 4% BSA in PBS (15 min at 37 C), and washed twice with PBS. IgE-sensitized BMMCs were left to attach to fibronectin-coated slides for 30 min in BSS-0.1% BSA at 37 C. Then, the cells were exposed to 50 μM UA or vehicle (0.1% DMSO) in BSS-0.1% BSA. After 15 min incubation at 37 C, the cells were exposed to AF 488-labeled anti-IgE for 10 min at 37 C in the presence of 50 μM UA or vehicle and then washed and fixed with 4% paraformaldehyde. Alternatively, the cells were incubated with UA or vehicle as aforementioned, washed, fixed, and then labeled for 10 min with AF 488-labeled anti-IgE. For Thy-1.1 mobility assay, RBL-2H3 cells were grown on sterilized 8-well multitest slides (MP Biomedicals) for 18 h, washed with BSS-0.1% BSA, and treated with anti-Thy-1.1 mAb. After 30 min, the antibody was washed out and the cells were exposed for 15 min at 37 C to 50 μM UA or vehicle. Then, the cells were exposed to AF 488-labeled anti-IgG for 10 min, washed, and fixed with 4% paraformaldehyde. As controls, anti-Thy-1.1-labeled cells were exposed to 50 μM UA or vehicle, fixed, and labeled with AF-labeled secondary antibody. The cells were washed and the samples were mounted using 90% glycerol supplemented with 5% n-propyl gallate. The mobility of FcεRI and Thy-1.1 was determined in the analytical volume comprising the apical part of the attached cells defined by pinhole size 1.5 AU. Images of the cells were acquired with a confocal laser scanning microscope Leica TCS SP5 equipped with X40/1.40 NA. oil-immersion objective. In each experiment, all images were acquired at identical microscope settings. The distribution of FcεRI and Thy-1.1 in individual cells was evaluated, and the fraction of fluorescence detected in the cell periphery (1 μm width) and centrally was calculated using ImageJ (106; https://imagej.nih.gov/ij/). Data are presented as a fraction of fluorescence localized centrally.

FRAP analysis
FRAP measurements were performed using the Leica Stellaris 8 Falcon microscopic system equipped with an oil immersion objective (60× /1.4 NA) and an environmental chamber controlling CO 2 concentration (5%) and temperature (37 C). The cells were grown in glass-bottomed wells of the 96-well plate (In Vitro Scientific, 96101N), and the photobleaching was achieved by a 100 ms laser pulse at 488 nm with 100% of the laser power level. Fluorescence intensity was monitored in a circular area of 1 μm in diameter. Confocal images were obtained from a single optical plane corresponding to the point of contact between the cell and the glass. The obtained time lapses were analyzed using LasX software (https://www.leica-microsystems.com/products/microscopesoftware/p/leica-las-x-ls/). Subsequently, the background was subtracted from the intensities of the FRAP area, and curves were corrected for bleaching. The resulting FRAP curves (15-18 measurements for each experiment) were fitted individually biexponentially to obtain an immobile fraction (107,108). All optimizations were performed in MATLAB R2021b (MathWorks), using a nonlinear leastsquares fit (lsqcurvefit).

Statistical analysis
Statistical significance of intergroup differences was determined by one-way ANOVA with Tukey's post-test using Prism version 5.04 graphics and statistic software package (Graph-Pad). Data in Figure 8, B and D were calculated with one-way ANOVA with Tukey's multiple comparison test. In experiments where only two groups were compared, the statistical significance of differences was evaluated with Student's unpaired, two-tailed t test. All values are expressed as means ± SEMs. Significance levels: *p < 0.05; **p < 0.01; ***p < 0.001.

Data availability
All data are contained within this article and in the supporting information.
Supporting information-This article contains supporting information.
Institute of Molecular Genetics of the Czech Academy of Sciences for help with analysis of FcεRI, Thy-1.1, and cholesterol mobility. This work was supported by project 20-16481S from the Czech Science Foundation, project FV20479 from the Ministry of Industry and Trade of the Czech Republic, project TP01010060 from the Technology Agency of the Czech Republic, project LM2018129, CZ.02.1.01/0.0/0.0/18_046/0016045 from the of the Czech Republic, and the Institutional project RVO 68378050 from the Czech Academy of Sciences. Conflicts of interest-The authors declare that they have no conflict of interest with the contents of this article.