Lysophosphatidic acid triggers mast cell-driven atherosclerotic plaque destabilization by increasing vascular inflammation.

Lysophosphatidic acid (LPA), a bioactive lysophospholipid, accumulates in the atherosclerotic plaque. It has the capacity to activate mast cells, which potentially exacerbates plaque progression. In this study, we thus aimed to investigate whether LPA contributes to plaque destabilization by modulating mast cell function. We here show by an imaging mass spectrometry approach that several LPA species are present in atherosclerotic plaques. Subsequently, we demonstrate that LPA is a potent mast cell activator which, unlike other triggers, favors release of tryptase. Local perivascular administration of LPA to an atherosclerotic carotid artery segment increases the activation status of perivascular mast cells and promotes intraplaque hemorrhage and macrophage recruitment without impacting plaque cell apoptosis. The mast cell stabilizer cromolyn could prevent intraplaque hemorrhage elicited by LPA-mediated mast cell activation. Finally, the involvement of mast cells in these events was further emphasized by the lack of effect of perivascular LPA administration in mast cell deficient animals. We demonstrate that increased accumulation of LPA in plaques induces perivascular mast cell activation and in this way contributes to plaque destabilization in vivo. This study points to local LPA availability as an important factor in atherosclerotic plaque stability.

Hereto, mice were sacrifi ced by perfusion through the left cardiac ventricle with phosphate-buffered saline (150 mM NaCl, 1.5 mM NaH 2 PO 4 , 8.6 mM Na 2 HPO 4 , pH 7.4) followed by perfusion with 128 mM ammonium bicarbonate to remove salts which interfere with secondary ion mass spectrometry (SIMS) measurements. Subsequently, the common carotid arteries were both excised, embedded in 10% gelatin at 30°C and snap-frozen in liquid nitrogen for optimal lipid preservation.
Transverse 10 µm cryosections were prepared on a Leica CM 3050 cryostat (Leica Microsytems, Rijswijk, The Netherlands) at Ϫ 20°C. The sections were cut in a proximal direction from the carotid bifurcation and mounted in order on a parallel series of glass slides, alternating between 1% gelatin-coated slides for immunohistological staining and conductive transparent indium tin oxide-coated slides (Delta Technologies, Stillwater, MN) for timeof-fl ight (TOF)-SIMS, which are stored at Ϫ 80°C until further use. The sample surface of a series of sections mounted on conductive slides were sputter coated with gold using a Quorum Technologies (Newhaven, East Sussex, UK) SC7640 sputter coater.
All static SIMS experiments were performed on a Physical Electronics (Eden Prairie, MN) TRIFT-II TOF-SIMS instrument described elsewhere ( 26 ) and newly equipped with a gold liquid metal ion gun. Secondary ions were extracted through a 3.2 keV electric fi eld into the TOF analyzer and postaccelerated by an additional 8 keV fi eld prior to detection on a dual multichannel plate/phosphor screen detector. All experiments were performed with a primary ion beam current of 1 nA, a primary pulse length of 18 ns, and a primary ion energy of 22 keV. The ion dose was such that all analyses were conducted at or under the static SIMS threshold (10 13

Cell culture
The capacity of LPA to activate murine mast cells was studied on MC/9 mast cells as well as on freshly isolated peritoneal mast cells (PMCs). MC/9 cells were cultured as described previously ( 6 ). Bone marrow-derived mast cells (BMMCs) were cultured by culturing bone marrow cells at a density of 0.25 × 10 6 cells in RPMI containing 10% fetal bovine serum (FBS), 2 mmol/l L -glutamine , 100 U/ml penicillin, 100 g/ml streptomycin, and murine IL3 for 4 weeks. Total RNA was extracted from these cells with guanidine thiocyanate , reverse transcribed using M-MuLV reverse transcriptase (RevertAid, MBI Fermentas, Leon-Roth, Germany) and expression of target genes was measured by quantitative PCR on an ABI PRISM 7500 Taqman apparatus (Applied Biosystems, Foster City, CA).
In this study, we have therefore investigated the potential involvement of LPA in mast cell-driven plaque destabilization. We demonstrate that LPA can activate adventitial mast cells, promote monocyte recruitment and microvascular leakage, and lead to an enhanced incidence of intraplaque hemorrhage. Finally, our data indicate that the adverse events of LPA-induced mast cell activation can be largely inhibited by the mast cell stabilizer cromolyn.

MS imaging
Lipid distribution of carotid artery plaques in LDLr Ϫ / Ϫ mice was determined at 9 weeks after perivascular collar placement.

Histology
Mast cells were visualized by staining of 5 m cryosections with aqueous Toluidine blue (Sigma). Neutrophils were stained with naphthol AS-D chloroacetate esterase (Sigma). Iron staining was performed according to Perl's method. Monocytes/macrophages were stained immunohistochemically with monocyte/macrophage antibody-2 (MOMA-2) (monoclonal mouse IgG2a, dilution 1:50; Serotec, Kidlington, Oxford, UK ). Apoptosis was visualized using a terminal deoxytransferase dUTP nick-end labeling (TUNEL) kit (Roche Diagnostics). Morphometric analysis (Leica QWin image analysis software) was performed on hematoxylin-eosin stained sections of the carotid arteries at the site of maximal stenosis. Toluidine blue-stained sections were used for histological examination for the presence of perivascular mast cells. Neutrophil numbers, mast cell numbers, the extent of mast cell degranulation, and the presence of iron were assessed manually. A mast cell was considered resting when all granula were maintained inside the cell, while mast cells were assessed as activated when granula were deposited in the tissue surrounding the mast cell. MOMA-2 and TUNEL-positive areas were quantifi ed with Leica QWin image analysis software and, in addition, TUNELpositive nuclei were counted manually. All morphometric analyses were performed by blinded independent operators.

Statistical analysis
Data are expressed as mean ± SEM. A two-tailed Student's t -test was used to compare individual groups. Non-Gaussian distributed data were analyzed using a Mann-Whitney U test. Frequency data analysis was performed by means of the Fisher's exact test. A level of P < 0.05 was considered signifi cant.

MS imaging
We have applied imaging mass spectrometry via TOF-SIMS for verifi cation and high-resolution spatial localization of lipids in carotid artery lesions from LDLr Ϫ / Ϫ mice. SIMS analysis shows strong signals for inorganic ions and ionized organic compounds derived from fragmentation of surface molecules. A number of LPA species could be detected in plaque material, and were tentatively assigned on the basis of their M+ peak ( Fig. 1A ). Flanking sections were stained for hematoxylin-eosin and macrophage content ( Fig. 1B, C ) to correlate intraplaque lipid distribution profi les to morphological features. Color intensity of the ion micrographs corresponded to signal strength ( ␤ -Hexosaminidase activity was determined by adding 50 l of releasate to 50 l 2 mM 4-nitrophenyl N -acetyl-b-D-glucosaminide (Sigma) in 0.2 M citrate (pH 4.5) and incubated at 37°C for 2 h. After addition of 150 l 1 M Tris (pH 9.0), absorbance (optical density) was measured at 405 nm (OD405). To measure chymase and tryptase release after degranulation, 50 l supernatant was added to 2 mM S-2288 (tryptase substrate, Chromogenix, Lexington, MA) or S-2586 (chymase substrate, Chromogenix) in PBS supplemented with 100 U/ml heparin. After 2 h (tryptase) or 24 h (chymase) at 37°C, OD405 was measured. Values are expressed as percentage of total release. MCP-1 levels were determined by ELISA according to the manufacturer's protocol (eBioscience).
RAW 264.7 macrophage cells were cultured in Dulbecco's modifi ed Eagle medium (DMEM) containing 10% FBS, 2 mmol/l L -glutamine, 100 U/ml penicillin, and 100 g/ml streptomycin. For proliferation experiments cells were seeded in 24-well dishes at a density of 1 × 10 5 cells/ml in DMEM containing 1% FBS, 2 mmol/l L -glutamine, 100 U/ml penicillin, and 100 g/ml streptomycin for 24 h to synchronize cell cycle. Subsequently, fresh serum-free DMEM containing 0.2% BSA and 18:1 LPA at various concentrations was added to the cells and cultured for 40 h without medium change. Ten percent FBS in DMEM was used as a positive control. After 16 h [ 3 H]thymidin (5.0 µCi/well; GE Healthcare, Eindhoven, The Netherlands) was added and cells were incubated further for 24 h . Thereafter, cells were washed three times with PBS, lysed with 0.1 mol/l NaOH, and cell-associated radioactivity was determined by liquid scintillation counting.
Cellular apoptosis was measured after overnight treatment of RAW 264.7 macrophages with tryptase (10 ng/ml, Sigma) in the presence or absence of LPA (22 µmol/l, 10 µg/ml). Cells were collected and stained with Annexin V/PI after which the percentage of cell death was measured using fl ow cytometry.

Microvascular leakage
Microvascular permeability was assessed as previously described ( 6 ). In short, male C57Bl/6 mice were injected intradermally at randomized sites with 5 × 10 5 MC/9 mast cells suspended in PBS containing 50 g/ml compound 48/80 or 18:1 LPA (10 µg/ml) (n = 6 per group). Immediately after intradermal injection of the cell suspensions, 100 l 1.25% Evans Blue was injected intravenously and after 30 min the surface area of Evans Blue-stained skin was measured. To measure MC/9 activation in skin, MC/9 cells using the conditions as described above were randomly injected intradermally in mast cell defi cient Kit(W Ϫ sh /W Ϫ sh ) mice, which lack endogenous skin mast cells. After 30 min, skin spots were excised and embedded in OCT compound for cryosectioning. Sections 10 µm thick were routinely stained with Alcian Blue/Safranin O (Sigma) and mast cell activation status was scored manually.

Local LPA challenge
Six weeks after collar placement, when advanced atherosclerotic lesions had developed, apoE Ϫ sh ) mice were challenged perivascularly by applying pluronic F-127 gel (25% w/v) or pluronic F-127 gel containing LPA (20 mol/l, ‫ف‬ 92 ng/animal) proximal to the collar (n = 8-12 per group). A subset of the LPA-challenged animals received an intravenous injection of the mast cell stabilizer cromolyn (25 mg/kg, Sigma) ( 6 ) thirty minutes before and twice daily after local challenge (50 mg/ kg ip). Three days after LPA-challenge mice were anesthetized and in situ fi xation through the left cardiac ventricle was performed, after which the carotid artery lesions were excised for further analysis .

Mast cell releasate
To determine whether LPA has the capacity to activate mast cells, releasates of compound 48/80 and LPA stimulated MC/9 cells, BMMCs, and freshly isolated peritoneal mast cells were analyzed. Supernatants of untreated mast cells were used as negative controls (0% release), while cell lysates of untreated cells were used for total mast cell content analysis (100%). Both compound 48/80 and LPA were able to induce release of ␤ -hexosaminidase activity and tryptase from MC/9 cells. ␤ -Hexosaminidase activity in the releasate of compound 48/80 and LPA stimulated MC/9 cells did not differ (3.1 ± 0.8% vs. 5.7 ± 1.5% of total release, respectively). Interestingly, tryptase release was much  that LPA 1 was the main LPA receptor expressed by our murine mast cell cultures (supplementary Fig. III), while LPA 2 and LPA 3 were undetectable. LPA 1 expression was highest in BMMCs that were skewed toward a connective tissue type mast cell, the predominant phenotype in the vessel wall ( 6 ).
The involvement of LPA 1 in LPA-mediated mast cell activation was further confi rmed by our PCR data, establishing plaque. We have also fi rmly established that LPA is a potent mast cell activator. To assess a functional role of LPA in adventitial mast cell activity and plaque progression in vivo, LPA (18:1) was applied perivascularly through a pluronic F-127 gel at the collar-induced carotid artery lesion in apoE Ϫ / Ϫ mice. As expected, morphometric analysis of the carotid lesions 3 days after challenge did not reveal any differences in plaque size between control and LPA-challenged animals (92 ± 13 × 10 3 m 2 vs. 96 ± 110 × 10 3 m 2 respectively, Fig. 3A ). Treatment with the mast cell stabilizer cromolyn for 3 days did not affect plaque size as well (93 ± 10 × 10 3 m 2 ). Additionally, no differences were found in medial surface area and percentage of artery stenosis (data not shown).
The number of adventitial mast cells as determined by toluidine blue staining did not differ between LPA-and mock-challenged mice ( Fig. 3B ), but the percentage of degranulated perivascular mast cells was still increased at time of sacrifi ce (three days after LPA stimulation, 53 ± 6% vs. 36 ± 7% activated mast cells in control animals, P = 0.04, Fig. 3C ). As mast cell activation occurs very acutely after application of a stimulus, it is expected that mast cell activation was even further increased earlier after application of LPA . Cromolyn treatment prevented the LPA-induced mast cell activation (38 ± 8% activated mast cells). Perivascular and intimal neutrophil numbers were unaltered after LPA challenge (data not shown). Interestingly, plaque macrophage content was seen to be enhanced upon LPA treatment (15.8 ± 2.7% vs. 9.6 ± 2.0% in control animals, P < 0.05, Fig. 4A ), and this effect did not occur in cromolyntreated mice (7.2 ± 1.3%, P < 0.01 vs. LPA) suggesting that the macrophage accumulation at least in part resulted from mast cell activation. This may possibly be caused by the LPA-mediated release of MCP-1 as shown in vitro. Furthermore, as LPA has been reported to promote macrophage survival ( 28 ), we examined the direct effect of LPA on a murine macrophage cell line. We show here that LPA induced a dose-dependent proliferative response in RAW 264.7 murine macrophages ( Fig. 4B ). No differences were seen with respect to elastic lamina breaks between in C57Bl/6 mice; PBS: 1.0 ± 0.12 vs. LPA: 1.08 ± 0.14 in Kit(W Ϫ sh /W Ϫ sh ) mice] and that of ␤ -hexosaminidase [PBS: ) mice], all displayed as fold change compared with PBS. Taken together, these data demonstrate that LPA is a potent inducer of LPA 1 -dependent mast cell activation, and in particular of tryptase and MCP-1 release.

Microvascular leakage in vivo
Mast cells are potent inducers of vascular leakage, which induces plaque destabilization as shown previously ( 6 ). To determine whether LPA-mediated mast cell activation enhances vascular leakage, we injected LPA and LPA-activated mast cells intradermally and quantifi ed leakage with Evans Blue dye. LPA and compound 48/80 per se were already seen to induce vascular leakage ( Fig. 2B , P < 0.05 and P < 0.001, respectively), due to activation of resident dermal mast cells as demonstrated previously ( 6 ). MC/9 cells without stimulus also induced vascular leakage, possibly due to shear stress-induced activation during injection. We observed that intradermal injection of LPA-activated MC/9 cells highly and signifi cantly induced vascular leakage ( P < 0.001 compared with PBS control, P < 0.05 compared with MC/9 only, Fig. 2B ) similarly as the positive control, compound 48/80 activated mast cells ( P < 0.001 compared with PBS ). Intradermal MC/9 injection in mast cell defi cient Kit(W Ϫ sh /W Ϫ sh ) mice unequivocally showed that compound 48/80 and LPA induced massive and persistent activation of transferred mast cells in skin ( Fig. 2C , P < 0.01 and P < 0.05, respectively, compared with unstimulated MC/9 cells). These data further establish that LPA is a potent mast cell activator which may affect atherosclerotic plaque stability in vivo.

Local LPA treatment and plaque morphology
We and others have previously demonstrated that LPA progressively accumulates in plaques during disease progression ( 20,21 ) and we here confi rmed the presence and actual location of a range of LPA species in the atherosclerotic Fig. 3. Adventitial mast cell content of atherosclerotic carotid artery lesions after focal LPA administration. A: Plaque size is not changed at three days after local LPA challenge as compared with control mice or mice that had received the mast cell stabilizer cromolyn. B: Total adventitial mast cell contents of control, LPA-treated, and LPA/cromolyn-treated mice were essentially similar. C: Adventitial mast cell degranulation was increased at three days after local LPA challenge as compared with control mice (* P < 0.05). Cromolyn treatment normalized the levels of mast cell activation in the LPA-challenged animals. challenge did not enhance the incidence of intraplaque hemorrhage and hemosiderin deposits in lesions in cromolyn-treated mice as these showed comparable numbers to the control group (20%, Fig. 4C ). Perl's iron staining confi rmed these fi ndings: 7 of 23 (35%) iron positive plaques in the LPA group compared with 2 of 20 (10%) in the control group ( Figs. 4D ).
To elucidate whether the observed effects of LPA in vivo are mast cell dependent, we challenged mast cell-defi cient apoE Ϫ / Ϫ Kit(W  Fig. 5A ). We did not observe signifi cant effects on intimal macrophage accumulation between the controls and LPA treated apoE Ϫ / Ϫ Kit(W  Fig. 5C ), groups (LPA: 2.1 ± 0.6 vs. 1.8 ± 0.5 in control animals and 1.4 ± 0.3 in LPA/cromolyn animals). Adventitial and in vitro mast cell degranulation was previously reported to promote apoptosis of vascular smooth muscle cells, endothelial cells, and macrophages ( 6,29 ). TUNEL staining, however, did not reveal any differences in the rate of intimal apoptosis three days after LPA challenge (1.5 ± 0.4% TUNEL positive nuclei vs. 1.2 ± 0.3% in the controls and 2.2 ± 0.6% in LPA/cromolyn animals). In addition, necrotic core size did not differ between the groups (data not shown). In vitro, LPA was not able to rescue macrophages from tryptase-induced apoptosis, suggesting that mast cell-induced apoptosis was compensated for by de novo infl ux of leukocytes, and possibly mediated by mast cell-derived MCP-1 (supplementary Fig. IV ).
Interestingly, LPA-challenged plaques demonstrated the presence of intraplaque hemorrhages characterized by accumulation of intimal erythrocytes in 45% of the plaques in LPA-challenged animals (10 of 23, Fig. 4C ), while intraplaque hemorrhage was observed in only 20% of the plaques in control mice (4 of 20, Fig. 4C ). The LPA plaque apoptosis was not altered upon LPA challenge, despite the preferential secretion of tryptase, previously shown by us and others ( 6,33 ) as main culprit in mast cellassociated apoptosis. Possibly the pro-apoptotic effects of the mast cell secretome are counteracted by the intrinsic histone deacetylase and serine threonine kinase (Akt)dependent anti-apoptotic effects of LPA itself ( 27,34 ). However, in our in vitro studies LPA was not able to rescue macrophages from tryptase induced apoptosis, rendering this theory unlikely. Hypothetically, LPA may also lower the capacity of monocyte-derived cells to emigrate from the vessel wall ( 35 ). LPA elicited effects on macrophage content were however completely abolished by cromolyn treatment, which was previously demonstrated to mainly target mast cells without any side effects on other cell types such as neutrophils and macrophages at the dosage used in our animal model ( 6 ). It is conceivable that chemotactic effects exerted by LPA-activated mast cells outbalance the aforementioned intrinsic LPA effects. It has already been demonstrated that murine mast cells release a plethora of chemokines including the CXCR2 ligand KC which can trigger directly and indirectly, via very late antigen-4 ( 36 ), leukocyte arrest to the endothelium. Furthermore, we now demonstrate that LPA favors mast cell release of MCP-1, which has long since been recognized as one of the primary cytokines in monocyte recruitment to the plaque ( 37 ). We thus hypothesize that mast cell degranulation promotes monocyte recruitment, while LPA on its own account will prolong their life span and inhibit their emigration, thereby increasing the plaque macrophage content. Mast cell stabilization by cromolyn possibly reverses this increase by preventing the fi rst step of monocyte recruitment and adhesion.
A few limitations of the study are worth mentioning. First, LPA is a rather heterogeneous class of monoacylated phospholipids that can impact a range of processes besides mast cell function. However, LPA-induced phenotypic which suggests that LPA induces intraplaque hemorrhage via a mast cell-dependent mechanism. DISCUSSION Activated mast cells were shown to accumulate in the arterial adventitia and intima during plaque progression and to promote plaque progression and plaque destabilization (4)(5)(6). Oxidized LDL has been proposed as the endogenous stimulus that may be responsible for mast cell activation in atherosclerosis ( 15,16 ). One of oxLDL's main lipid constituents, LPA, has a wide range of effects on mast cells, infl uencing processes such as proliferation, differentiation, and release of histamine, macrophage infl ammatory protein-1 ␤ , IL-8, eotaxin, and MCP-1 ( 18,19 ). Moreover, as this bioactive lipid progressively accumulates in plaques during disease progression ( 20,21 ), it may be a likely candidate for vascular mast cell activation in the context of atherosclerosis.
Concordant with our earlier fi ndings, we were able to detect several LPA species in mouse plaques, in particular in the intimal lipid core of the plaque underlying the fi brous cap of the lesion. Our in vitro and in vivo experiments not only confi rm that LPA is able to induce mast cell activation but also demonstrate that the secretion pattern after LPA challenge, with preferential release of tryptase over ␤ -hexosaminidase and chymase, differs from that of other stimuli such as compound 48/80, indicating that different activators have different mast cell releasate profiles. In addition, we demonstrate here that LPAinduced mast cell activation enhances microvascular leakage, allowing the infl ux of detrimental agents and hematopoietic cell subsets such as monocytes and erythrocytes into the atherosclerotic plaque.
One of the features of LPA treatment indeed appeared to be the acute increase in intraplaque hemorrhage incidence, a phenomenon that was also observed after dinitrophenyl hapten-elicited mast cell activation in situ ( 6 ). Intraplaque hemorrhage is a potent pro-atherogenic stimulus and risk factor in plaque destabilization, as it is accompanied by deposition of erythrocyte-associated cholesterol and enlargement of the necrotic core of the atherosclerotic plaque. Treatment of mice with cromolyn during LPA challenge, similarly as after dinitrophenyl hapten challenge ( 6 ), normalized the extent of mast cell degranulation in the adventitia, while largely preventing intraplaque hemorrhage, implying that mast cells are directly involved in LPA-mediated plaque destabilization .
During the development of atherosclerosis, tissue concentrations of LPA increase ( 21 ), thus enhancing local bioavailability of LPA with the capacity to destabilize plaques. Apoptosis at later stages of disease progression is considered deleterious for plaque stability ( 30 ). As mast cell degranulation was reported to promote apoptosis of smooth muscle cells ( 31,32 ), endothelial cells ( 29 ), and intimal macrophages ( 6 ), we established the intimal apoptotic nuclei content after LPA challenge. To our surprise, changes in the plaque observed in this study did not occur after treatment with the mast cell stabilizer cromolyn and in mast cell defi cient apoE Ϫ / Ϫ Kit(W Ϫ sh /W Ϫ sh ) mice which suggests that in this experimental setup the mast cell is a prominent mediator of the LPA-induced plaque destabilization. Second, the LPA dose used in this study may far exceed endogenously available LPA levels, leading to nonphysiological responses. In previous studies however, we demonstrated that in advanced lesions the level of LPA is approximately 40 pmol/mg atherosclerotic tissue, which corresponds to approximately 35 ng of LPA per plaque. In this study, we locally applied 92 ng of LPA in a pluronic gel that releases its content within three days, which suggests that the dose of LPA used is within a physiological range. Third, in this study we topically applied 18:1 LPA, while in plaque a number of polyunsaturated LPA species were found to accumulate. The latter were shown to display an even higher potency in, for example, platelet activation ( 38 ) and this may also hold true for mast cell activation. However, 18:1 LPA is one of the most abundant LPA species in plaque, rendering it highly likely that it is directly accessible for plaque mast cells.
In conclusion, we here provide in vivo proof that LPA bioavailability in the plaque is an important factor in atherosclerotic plaque stability and that effects of LPA are at least partly mediated by LPA-induced mast cell activation. Moreover, we show that enhanced LPA levels in or close to the plaque, as occurring during plaque progression, will increase the plaque macrophage content and induce vascular leakage in a mast cell-dependent manner. LPA has previously been shown to have a number of thrombogenic and atherogenic actions as also recently reviewed by Schober and Siess ( 39 ) and these data provide yet another mechanism by which LPA may affect atherosclerotic plaque progression and destabilization. We propose that intervention in LPA bioavailability or activity in the plaque may be an effective measure to reduce mast cell activation and vascular infl ammation, improve plaque stability, and concomitantly reduce plaque thrombogenicity, and thus could well represent an effective therapeutic strategy in the prevention of acute coronary syndromes.