Platelet activation independent of pulmonary inflammation contributes to diesel exhaust particulate-induced promotion of arterial thrombosis

Background Accelerated thrombus formation induced by exposure to combustion-derived air pollution has been linked to alterations in endogenous fibrinolysis and platelet activation in response to pulmonary and systemic inflammation. We hypothesised that mechanisms independent of inflammation contribute to accelerated thrombus formation following exposure to diesel exhaust particles (DEP). Methods Thrombosis in rats was assessed 2, 6 and 24 h after administration of DEP, carbon black (CB; control carbon nanoparticle), DQ12 quartz microparticles (to induce pulmonary inflammation) or saline (vehicle) by either intra-tracheal instillation (0.5 mg, except Quartz; 0.125 mg) or intravenous injection (0.5 mg/kg). Thrombogenicity was assessed by carotid artery occlusion, fibrinolytic variables and platelet-monocyte aggregates. Measures of inflammation were determined in plasma and bronchoalveolar lavage fluid. Tissue plasminogen activator (t-PA) and plasminogen activator inhibitor (PAI)-1 were measured following direct in vitro exposure of human umbilical vein endothelial cells (HUVECs) to DEP (10–150 μg/mL). Results Instillation of DEP reduced the time to thrombotic occlusion in vivo, coinciding with the peak of DEP-induced pulmonary inflammation (6 h). CB and DQ12 produced greater inflammation than DEP but did not alter time to thrombotic occlusion. Intravenous DEP produced an earlier (2 h) acceleration of thrombosis (as did CB) without pulmonary or systemic inflammation. DEP inhibited t-PA and PAI-1 release from HUVECs, and reduced the t-PA/PAI-1 ratio in vivo; similar effects in vivo were seen with CB and DQ12. DEP, but not CB or DQ12, increased platelet-monocyte aggregates. Conclusion DEP accelerates arterial thrombus formation through increased platelet activation. This effect is dissociated from pulmonary and systemic inflammation and from impaired fibrinolytic function. Electronic supplementary material The online version of this article (doi:10.1186/s12989-016-0116-x) contains supplementary material, which is available to authorized users.


Background
The association between air pollution and cardiovascular disease is well-established [1], with exposure to trafficderived pollutants implicated in acute myocardial infarction [2,3]. The adverse cardiovascular effects of airborne pollution have been attributed to combustion-derived particulate matter (PM), especially ultrafine particles from diesel exhaust (DEP), that has a large reactive surface area and can penetrate deep within the lung (reviewed in [4]). Recent investigations suggest that DEP [5] and ambient particulate matter [6,7] increase cardiovascular events by accelerating thrombosis [2].
The mechanisms responsible for PM-induced enhancement of thrombosis remain controversial. Diesel exhaust inhalation in humans produced endothelial cell dysfunction 2-24 hours after exposure [8][9][10], accompanied by impaired endogenous fibrinolysis [8], increased thrombogenicity and platelet activation [5]. These effects could be directly attributed to the exhaust particles as they were markedly reduced following removal of particles from the exhaust [11,12]. Investigations using appropriate animal models support the contention that DEP increases the likelihood of thrombus formation [13,14]. However, whilst several different mechanisms for enhanced coagulation in response to particles have been implicated (including impaired fibrinolysis [7] and platelet activation [7,15,16]), there remains much disagreement. A recent study in mice, indicated that increased thrombus formation could be attributed to PM-induced pulmonary and systemic inflammation via multiple mechanisms [7]; including an increased tumour necrosis factor (TNF)-αdependent impairment of fibrinolysis and an interleukin (IL)-6-dependent activation of coagulation. However, whether the cardiovascular effects of DEP require direct interaction with the lung is uncertain since intravenous (i.v.) injection caused systemic and pulmonary inflammation associated with alterations in the cardiovascular system in spontaneously hypertensive rats [17].
There is also some evidence that DEP can, at least acutely, increase coagulation independent of pulmonary inflammation. Inhibition of pulmonary inflammation 1 h after exposure of hamsters to DEP had no effect on coagulation or platelet activation, whereas the same intervention did reduce DEP-mediated enhancement of coagulation 3 and 24 h after exposure [18]. It was suggested that penetration of DEP, or components of DEP, into the blood could account for this early pro-thrombotic effect. This would certainly be consistent with the demonstration that direct exposure of blood to DEP can enhance thrombus formation ex vivo [1].
Whilst previous investigations have considered the role of pulmonary inflammation, uncertainty remains about its influence on the pro-thrombotic actions of DEP. The current investigation addressed the hypothesis that exposure to DEP induces an increase in thrombosis independent of pulmonary inflammation. This was addressed in vivo by using (1) intra-pulmonary and (2) intravenous administration of DEP or appropriate 'control particles' (carbon black, DQ12 quartz). These studies were complemented by in vitro experiments to determine whether direct exposure to DEP alters endogenous fibrinolysis in endothelial cells.

IT instillation of particles
IT instillation (500 μg DEP/rat) resulted in successful administration of DEP, with particle aggregates evident in alveoli (but not in liver, kidney or carotid artery thrombi) collected 2 or 6 h after administration (data not shown).

Pulmonary & systemic inflammation
IT instillation of DEP, CB or DQ12 quartz induced clear, transient pulmonary inflammation with a marked influx of cells (predominantly neutrophils), peaking 6 h post instillation (Fig. 1). The responses produced by CB and DQ-12 particles were greater than those produced by DEP (Fig. 1). Cellular infiltration was accompanied by increased concentrations of interleukin (IL)-6 in bronchoalveolar lavage fluid (BALF) 6 h after instillation (Fig. 2i), although apparent increases in tumour necrosis factor (TNF)-α (Fig. 2ii) and C-reactive protein (CRP; Fig. 2iii) did not achieve statistical significance. The indication that the pulmonary inflammation produced by control particles was more marked than that induced by DEP was supported by the demonstration that CB increased both IL-6 ( Fig. 2i) and CRP (Fig. 2iii), whilst DQ12 quartz increased CRP concentrations in BALF (Fig. 2iii). Furthermore, control particles induced systemic inflammation: concentrations of CRP in plasma (6 h) were increased by IT instillation of CB or DQ12 quartz, but unaltered after IT instillation of DEP (Fig. 2v). Plasma concentrations of IL-6 and TNFα were not affected by IT instillation of any of the particles (data not shown).

Thrombus formation
FeCl 3 produced thrombus formation, leading to total occlusion (cessation of blood flow) of the carotid artery after 12-15 min in vehicle-treated rats. IT instillation of DEP (Fig. 3i) produced a pattern of reduced time-to-occlusion at all three time-points assessed (2, 6 & 24 h), with the difference achieving significance 6 h after exposure (DEP 9.4 ± 0.8 min vs vehicle 13.8 ± 0.7 min; n = 6). In contrast, neither CB nor DQ12 quartz accelerated thrombus formation (Fig. 3ii).

Endogenous fibrinolysis
Plasma PAI-1 and t-PA concentrations were measured to determine whether altered endogenous fibrinolysis could account for the accelerated thrombosis induced by DEP. These measurements demonstrated that IT instillation of DEP reduced plasma t-PA and increased PAI-1 concentrations, reducing the t-PA:PAI-1 ratio (Fig. 3(iii)). Instillation of CB or DQ12 quartz produced a similar impairment in fibrinolytic function ( Fig. 3(iii)).

Platelet-monocyte aggregation
Platelet-monocyte aggregation was increased 6 h after IT instillation of DEP ( Fig. 3(iv)). In contrast, instillation of CB or DQ12 Quartz had no effect.

IV injection of particulate
Successful administration of particles to the circulation was confirmed by the identification of DEP aggregates in sections of liver (although not in kidney) 2 h after intravenous injection (data not shown). No particles were identified in these organs 6 or 24 h after injection.

Absence of pulmonary and systemic inflammation
In contrast to IT instillation, i.v. injection of DEP or CB did not cause pulmonary inflammation. There was no influx of inflammatory cells into the lung (Fig. 4i), and no change in the inflammatory cell populations ( Fig. 4ii & iii) or the cytokine concentrations in BALF: TNFα (Additional file 1: Figure S1i), CRP were identified by nuclear morphology in haematoxylin and eosin-stained cytospins after in vivo administration of diesel exhaust (DEP; 2, 6 or 24 h) or control (carbon black (CB); DQ12 quartz; 6 h) particle solutions. DEP and CB were associated with inflammatory cells (arrows) or phagocytosed by macrophages (arrow heads). There was no evidence of phagocytosis of quartz particles. Magnification ×400. (ii) Instillation of DEP (0.5 mg/rat; black columns) caused a transient increase in total cell number that peaked 6 h following administration. CB (0.5 mg/rat; light grey columns) or DQ12 quartz (0.125 mg/rat; dark grey columns) also increased cell number 6 h after instillation. (iii) Increases in total inflammatory cell number were due to an increase in neutrophils. Data are mean ± s.e.mean (n = 6) and were compared with appropriate control using Student's unpaired t-test. *P < 0.05, ***P < 0.001 (Additional file 1: Figure S1ii). IL-6 concentrations were below the limit of detection (data not shown). In addition, there was no evidence of systemic inflammation as plasma concentrations of CRP, IL-6 and TNFα were unaffected by i.v. DEP. Interestingly, plasma CRP (but not IL-6 or TNFα; data not shown) was increased 2 h after injection of CB (Additional file 1: Figure S1iii).

Thrombus formation
In the FeCl 3 -induced thrombosis model, DEP (0.5 mg/ kg, i.v.) produced a transient reduction in time-toocclusion 2 h following administration, which was no longer evident after 6 or 24 h (Fig. 5ii). The control particles, CB, also produced a small reduction in timeto-occlusion at the 2 h time-point (Fig. 5ii). No Fig. 2 Influence of particle instillation on pro-inflammatory cytokines. Instillation of diesel exhaust particles (DEP) increased IL-6 concentrations in bronchoalveolar lavage fluid (BALF) (i), but apparent increases of tumour necrosis factor (TNF)-α (ii) and C-reactive protein (CRP) (iii) did not achieve significance. Carbon black (CB) increased concentrations of IL-6 and CRP, whilst DQ12 quartz increased CRP. (iv) In plasma, carbon black (CB; light grey columns) or DQ12 quartz (dark grey columns), increased concentrations of CRP compared with controls (white columns), whereas DEP (black columns) did not. Data are mean ± s.e.mean (n = 6) and were compared with relevant control using Student's unpaired t-test or one-way ANOVA, as appropriate. *P < 0.05, ***P < 0.001. ns = not significant particles were identified by histological analysis of the thrombosed vessel.

In vitro exposure of HUVECs to DEP
In order to determine whether DEP can directly alter the fibrinolytic system in endothelial cells, cultured human umbilical vein endothelial cells (HUVECs) were exposed (16 h) to DEP (10-150 μg/mL) or vehicle, with or without thrombin (1.0 U/ml, 24 h) stimulation.
Exposure to DEP was not cytotoxic as treatment had no effect on cell viability as measured by LDH assay (Additional file 2: Figure S2). Pre-exposure to DEP reduced expression of t-PA (Fig. 6i) and PAI-1 (Fig. 6ii). Exposure to DEP (150 μg/ml; 6-24 h) also reduced t-PA antigen concentrations (Fig. 6iii) and PAI-1 activity ( Fig. 6iv) in culture media. Stimulation with thrombin had no effect on t-PA antigen levels ( Fig. 6v) but increased PAI-1 activity (Fig. 6vi) in media, an effect that was attenuated in the presence of DEP.

Discussion
This investigation addressed the hypothesis that exposure to DEP induces accelerated thrombus formation independent of pulmonary inflammation. This was achieved by comparing the effects of pulmonary and systemic DEP administration with appropriate control particles (CB, a "clean" carbon particle to control for the nano-carbon core of DEP and DQ12 quartz, a large pro-inflammatory particle that will not translocate into the blood [19,20]). Instillation and injection of DEP (and other nanoparticles) have been shown to increase pulmonary and systemic inflammation, with circulating levels of IL-6, TNFα and leukotriene-B4 elevated in the rat [14,[20][21][22][23][24]. The current investigation demonstrated that control particles caused pulmonary and/or systemic inflammation but did not promote thrombosis. In contrast, DEP-induced acceleration of thrombus formation was evident in the absence of pulmonary or systemic inflammation. The outcomes of these experiments are summarised in Table 1. (ii) Control particles (CB (0.5 mg/rat) or DQ12 quartz (125 mg/rat;)) did not alter time to occlusion 6 h after instillation. (iii) The ratio of tissue plasminogen activator (t-PA) to plasminogen activator inhibitor (PAI)-1 was reduced in the plasma of rats following intra-tracheal (IT) instillation of DEP, CB or DQ12 quartz. (iv) IT instillation of DEP, but not CB or DQ12 quartz increased platelet-monocyte interactions. Data are mean ± s.e.mean (n = 6) and were compared with relevant controls using one-way ((ii, (iii), (iv)) or two way (i) ANOVA. *P < 0.05, **P < 0.01 and ***P < 0.001

Inflammation and accelerated thrombosis following intra-tracheal instillation of DEP
Intra-tracheal (IT) instillation has been used extensively to investigate the effects of pulmonary exposure to a number of different particles [13,24,25]. It provides excellent alveolar penetration and distribution, and allows direct comparison of different particle types, with increased thrombus formation and platelet activation reported after pulmonary instillation of silica [26], titanium dioxide (TiO2) nanorods [27], ambient particulate matter [14], and carbon nanotubes [28]. Instillation has the advantage of ensuring direct administration of a controlled particle dose, whereas with inhalation there can be uncertainty about the proportion of particles that actually reaches the lungs (due to retention in the nasal cavity) [29]. IT instillation or inhalation of combustionderived particles produce similar effects on thrombus formation in vivo [7]. The dose of DEP used for IT instillation (500 μg DEP/rat) is comparable with previous studies in animals (5-500 μg/animal for hamsters [13,21,24]) exploring particle activity, and can be used to consistently induce pulmonary inflammation.
The accelerated thrombosis induced 6 h after IT instillation of DEP was consistent with the pro-thrombotic effect described both in clinical [5,[8][9][10], and in animal (hamsters [13,21,30], mice [14,22], or rats [23,31]) studies. This enhanced thrombus formation was co-incident with the peak in pulmonary inflammation. Whilst this would be consistent with thrombosis increasing as a consequence of pulmonary inflammation [7], neither CB nor (iii) Differential analysis confirmed that the majority of cells were macrophages, both in saline-treated (white columns) and in particle-treated rats. Data are mean ± s.e.mean (n = 6). All comparisons with saline-treated control were statistically non-significant; P > 0.05 using two-way ANOVA (DEP vs Saline) or Student's unpaired t-test CB vs Saline) DQ12 quartz (which both induced inflammation) accelerated thrombus formation. This contrasts with the relationship between increased thrombosis and pulmonary inflammation reported for a number of different particles (silica [26], TiO2 [27], carbon nanotubes [28], DEP [13,22]). More recent evidence, however, has suggested that CB-induced thrombogenic effects are independent of pulmonary and systemic inflammation [32].
The possibility that increased thrombosis is the result of impaired endogenous fibrinolysis has been indicated in a number of investigations [3]. Endothelial t-PA release was reduced in healthy volunteers [8] and in men with stable coronary heart disease [9] 2-6 hours after exposure to DE. The particulate constituents of these emissions contributed to the impaired t-PA release [11]. Impaired endogenous fibrinolysis has also been reported in some (but not all [33,34]) studies in animals, with increased PAI-1 and reduced tissue factor pathway inhibitor following exposure to urban particulate matter [7,14,35]. The timing of these changes suggests involvement of an inducible pathway or changes in protein synthesis. Whether these changes are secondary to effects in the lung or are a consequence of direct interaction of DEP with the arterial wall has not been determined. The presence of pulmonary and systemic inflammation following particle administration may be expected to influence fibrinolysis since chronic inflammation is associated with impairment of the endogenous fibrinolytic system in vivo [36] and in cultured endothelial cells [37]. Furthermore, reduction of systemic inflammation (with the antioxidant ascorbic acid) restored fibrinolytic function in chronic smokers [38]. However, whereas our demonstration that IT instillation of DEP alters the endogenous fibrinolytic system is consistent with the impaired fibrinolysis reported after inhalation of DE in humans [8], the use of control particles indicated that altered fibrinolysis could not be the sole cause of increased thrombosis following instillation. CB or DQ12 quartz, which produced a similar impairment in fibrinolytic function, did not alter thrombus formation. These results suggest, therefore, that altered endogenous fibrinolysis may be just one component of DEPmediated acceleration of thrombosis. Interestingly, although pulmonary inflammation induced by urban particulate matter can alter endogenous fibrinolysis (increased PAI-1) through a TNF-α-mediated process, it activates coagulation via a distinct IL-6-mediated pathway [7,39,40]. The mechanism responsible for DQ12induced alteration of fibrinolytic function has not been identified, but may be related to the ability of particulate (iv) Platelet-monocyte interaction was increased following IV administration of DEP but not CB. Data are mean ± s.e.mean (n = 6) and were compared with relevant controls using two way ANOVA (DEP vs Saline) or Student's unpaired t-test (CB vs Saline). *P < 0.05, **P < 0.01, ***P < 0.001 exposure to stimulate cytokine release from inflammatory cells [41] or release of an as yet unidentified mediator into the circulation [42].
Since our data suggest that alteration of endogenous fibrinolysis cannot solely account for acceleration of thrombosis, we addressed the role of platelet activation. Instillation of polystyrene particles increased platelet activation [43] and platelet-granulocyte aggregation in experimental animals [14]. The increase in plateletmonocyte aggregation 6 h after IT instillation of DEP in our investigation coincided with accelerated thrombus formation. This is consistent with previous work [3], and with a clinical study from our group [9]. In contrast, exposure to CB or DQ12 quartz had no effect, paralleling the pattern of thrombus effects across these particles. Thus, in contrast to inflammation and altered fibrinolysis, platelet activation was only seen following exposure to the particle (DEP) that accelerated thrombosis.

Accelerated thrombosis without inflammation following i.v. injection of DEP
In addition to the role of inflammation, the cardiovascular consequences of particle exposure have also been Fig. 6 Diesel exhaust particles alter the endogenous fibrinolytic system in HUVECs. Human umbilical vein endothelial cells were exposed to diesel exhaust particles (DEP; 100 μg/ml, 16 h, black columns) or vehicle (white columns) then incubated in the presence or absence of thrombin (1.0 U/ml, 24 h). Pre-exposure to DEP reduced expression of (i) tissue plasminogen activator (t-PA) and (ii) plasminogen activator inhibitor (PAI-1). Exposure to DEP (150 μg/ml; 6-24 h) also reduced (iii) t-PA antigen concentrations and (iv) PAI-1 activity in culture media. Stimulation with thrombin did not significantly increase t-PA antigen levels in media (v) but increased PAI-1 activity (vi). Thrombin-stimulated PAI-1 activity was suppressed in cells pre-incubated with DEP. Data are mean ± s.e.mean (n = 6). Comparisons were made using Student's unpaired t-test; **P < 0.01, ***P < 0.001 attributed variously to alteration of the autonomic nervous system, translocation of particles from the lung into the systemic circulation, and oxidative stress [3]. Particle translocation has attracted attention since direct (i.v.) administration of particles to the blood increases thrombosis in vivo [13]. This may be due to direct effects on components of the cardiovascular system since ex vivo administration of DEP to blood accelerates clot CB carbon black, CRP C-reactive protein, DEP diesel exhaust particles, DQ12 quartz particles, HUVEC human umbilical vein endothelial cells, IL-6 interleukin 6, PAI-1 plasminogen activator inhibitor, PMA platelet monocyte aggregation, TNFα tumour necrosis factor-α, tPA tissue plasminogen activator NSD = no significant difference.
---= not tested (note, DQ12 was not tested by iv injection or cultured endothelial cells as this particle is unlikely to cross from the lungs into the circulation due to its micrometre size) formation [13]. We addressed the proposal that DEP could accelerate thrombus formation by direct interaction with the cardiovascular system using i.v administration of DEP (0.5 mg/kg) or control (CB) particles. The dose selected was based on previous i.v. exposures in rats (0.01-0.5 mg/kg; [21,23,31]. As expected i.v. injection of particles did not cause pulmonary inflammation. Enhanced thrombus formation 2 h after i.v. injection of DEP is consistent with the proposal that this response is not secondary to inflammation. Interestingly, injection of CB (which also accelerated thrombus formation) did increase plasma CRP adding to the evidence that CB is not an "inert" control particle [41,44]). This suggests that the nanocarbon core of injected DEP contributes to its pro-thrombotic action. This may be through direct interaction with the endothelium or generation of free radicals [45]. Certainly, CB (which models the nanocarbon core of DEP) alters endothelial cell function [41,44], including inhibiting t-PA expression and protein production in cultured HUVECs [46]. The reduced t-PA:PAI-1 ratio observed after injection of DEP or CB suggests that alterations in endogenous fibrinolysis contribute to accelerated thrombosis following i.v. administration.
The use of cultured HUVECs demonstrated that DEP can alter endogenous fibrinolysis through direct interaction with the vascular endothelium. In the current investigation there was a pattern for lower thrombininduced release of t-PA in cultured endothelial cells after exposure to DEP but this did not achieve significance [47]. However, our results are consistent with the suggestion that DEP can alter the endogenous fibrinolytic system by direct interaction with endothelial cells. Whether this occurs in vivo is debatable since the concentrations of DEP used were high and unlikely to be experienced by endothelial cells following pulmonary or systemic administration.
Particle administration to the blood in vitro directly activates platelets [5,13,43]. Since nanoparticles may translocate into the systemic circulation [48], it remains possible that inhaled particles could interact directly with platelets to accelerate thrombosis. In support of this notion, a reduction in platelet number (6 h) following i.v. injection of DEP in rats has been attributed to removal of aggregated platelets from the circulation [23]. Furthermore, the pro-thrombotic response to DEP has been linked to changes in platelet function [11,13,21,49]. As with IT instillation, i.v. injection increased plateletmonocyte interaction, co-incident with accelerated thrombosis. Because both DEP and CB have been shown to be internalised by platelets [50], the ability of DEP to activate platelets is likely to be mediated by chemical species on the surface of DEP. Further studies are required to more precisely determine the mechanisms of DEP-induced platelet activation, including the role of platelet surface makers and glycoproteins [49,50], interaction with circulating leucocytes and the blood vessel wall [51], changes in catecholamines [52] and the alterations in other blood constituents in susceptible models [53,54]. Overall, these findings suggest that platelet activation, independent of pulmonary or systemic inflammation, contributes to the accelerated thrombus formation induced by DEP.

Limitations
Identifying the mechanisms responsible for particleinduced acceleration of thrombus formation has proved challenging despite considerable research effort. The current study suggests a mechanism independent of systemic and pulmonary inflammation. It was designed to compare DEP suitable control particles, showing that particle induced inflammation does not necessarily increase thrombosis. As with all investigations, however, consideration must be given to potential limitations. An alternative approach to addressing this mechanism would be to use anti-inflammatory drugs or to study DEP-induced thrombosis in transgenic animals with impaired inflammatory pathways. The doses/exposures used in the experiments described here are higher than would be expected in the real world and are designed to provide conceptual insight into the mechanisms involved. They are also, however, single exposures in healthy animals; it is possible that chronic exposure of lower doses could be detrimental and may also have more effect in patients with underlying cardiovascular and/or respiratory disease. Furthermore, the magnitude of the effects produced is striking, suggesting that even lower levels could produce prominent effects in the blood. Blood thrombogenicity plays a key role in cardiovascular health and small imbalances towards increased clotting are associated with marked increases in the risk of cardiovascular events in patients with coronary artery disease and stroke.

Conclusions
This study has used comparison with control particles and different routes of administration to demonstrate that DEP-mediated acceleration of thrombosis in vivo is independent of systemic and pulmonary inflammation. DEP directly altered components of the endogenous fibrinolytic system in endothelial cells but these changes alone could not account for the acceleration of thrombus formation. In contrast, platelet activation in response to DEP did not require pulmonary or systemic inflammation and may play an important role in increased thrombus formation. Although nanoparticle translocation is possible [48,[55][56][57], it seems unlikely that it would result in sufficient exposure to account for the acute acceleration of thrombus following instillation or inhalation [58]. Furthermore, similar changes in fibrinolytic function have been reported following pulmonary administration of DQ12 quartz, a particle considered too large to translocate [19,20]. It is possible that this DEP-induced increase in thrombosis contributes to the increased cardiovascular disease risk in individuals exposed to vehicle emissions but whether this is the case remains to be determined.

Particle suspensions
Diesel exhaust particles (DEP; 1 mg/mL; Standard Reference Material 2975; National Institute of Standards and Technology (NIST) Gaithersburg, MD, USA), nano-Carbon Black (CB; 1 mg/mL; Printex90, Degussa, Germany) and quartz microparticles (0.25 mg/mL; DQ12, IUF, Germany) were suspended in sterile 0.9 % saline (in vivo experiments) or culture medium (EGM-2; Lonza Walkersville, USA; in vitro experiments) and dispersed by sonication (5 min; 70 % power, 5 cycles/s; Status Homogenisers, Philip Harris Scientific, UK). Suspensions were prepared immediately before use and were endotoxin free (Limulus Amebocyte Lysate assay, Cambrex, USA; [41]; data not shown). CB lacks the organic and metal constituents of DEP and represents a 'clean' carbon particle of a similar primary particle size to DEP. DQ12 quartz particles induce pulmonary inflammation but do not translocate from the lung due to their large particle diameter (~1 μm; [20]).

Impact of particle administration on thrombosis in vivo
Animal work was performed under Home Office Licence (UK) in accordance with the Animals (Scientific Procedures) Act, 1986 and following local ethical committee review. Adult male Wistar Rats (Charles River, UK; 175-275 g) were housed under standard light cycle (lights on 08:00-20:00) and temperature (21-22°C).

Carotid artery thrombosis
Arterial thrombus was induced by application of ferric chloride (FeCl 3 ; adapted from [59]). Briefly, carotid artery blood flow was measured in anaesthetised (isoflurane) rats using an ultrasonic flow probe (Transonic Systems Inc, Netherlands). FeCl 3 solution (20 %) was applied topically to the carotid artery (10 min), arterial blood flow monitored, and the time to cessation of blood flow recorded.

Tissue retrieval
Rats were killed by exsanguination. Blood was collected from the abdominal aorta, citrated (10:1 in 3.8 % sodium citrate solution), and either analysed immediately (flow cytometry) or stored as plasma. Tissues (carotid artery, kidney, liver, lung) were processed for histological analysis. Serial sections (4 μm thick) were stained with hematoxylin and eosin.

Pro-inflammatory cytokines in the lungs and systemic circulation
Plasma and BALF were assessed for inflammatory markers (interleukin (IL)-6, C reactive protein (CRP) and tumour necrosis factor alpha (TNFα)) by ELISA (R&D Systems, UK) to manufacturer's instructions, using an automated Triturus ELISA analyser (Grifols, Spain).

Endogenous fibrinolysis
Plasma samples were assessed for t-PA and active PAI-1 using commercially-available ELISA kits (Patricell Ltd, UK).

Platelet-monocyte aggregation
Platelet-monocyte aggregates were measured by flow cytometry using fluorescein isothiocyanate (FITC)-labelled anti-rat CD42d antibody [60] and phycoerythrin (PE)-labelled hamster anti-rat CD61 [61]. PE-labelled hamster IgG1κ was used as a negative isotype control (antibodies from BD Biosciences, UK). Samples were analysed by BD FACscan flow cytometer equipped with CellQuest software for data acquisition and FlowJo software for data analysis. Platelet-monocyte aggregates were defined as double positive for CD42d and CD61.