Three different studies were undertaken to study the effects of inflammation on tissue-specific hypercoagulability: 1. Development of a new method to assess TF and TM activity in tissue homogenates. In the first study a new method was developed to determine the activities of TF and TM within tissue homogenates. To this purpose, a thrombin generation assay was modified to allow the addition of tissue homogenates. 2. Validation of the new method using a mouse endotoxemia model. The new developed method to determine the activities of TF and TM within tissue homogenates was validated using a mouse LPS-induced endotoxemia model. 3. Effects of particulate matter on tissue factor activity and thrombin generation. The inflammation-driven procoagulant effects of particulate matter exposure on lung tissue were studied using a spontaneously hypertensive rat model.

C57Bl/J6 mice (Charles River, Someren, The Netherlands) were used as wild type (WT) control mice as well as in the LPS induced endotoxemia model. TM^pro/pro mice carry a single amino acid substitution (Glu404>Pro) in the TM protein, which causes a reduced expression of TM and disrupts the TM-dependent activation of protein C for almost 90% 19.

Endotoxemia was induced by intraperitoneal (i.p.) injection of 2 mg LPS (E. coli serotype O55:B5; Sigma-Aldrich, St. Louis, MO) per kg bodyweight. After 6 hours, mice were anaesthetized with 4% isoflurane and blood was drawn from the vena cava after injection of 7 % of the bodyweight 3.2% citrate buffer as described previously 20. Mice were perfused using St. Thomas cardioplegic solution (University Hospital Maastricht, The Netherlands) through the portal vein for 10 minutes. Lung, liver, heart, kidney, brain, and spleen were snap frozen in liquid nitrogen. All mice were 10–12 weeks old and n = 6 mice per group were used. All mice studies were approved by the Animal Care and Use Committee of the Maastricht University.

Eleven to twelve weeks old spontaneously hypertensive (SHR) male rats were exposed to PM by intratracheal instillation 21. In brief, rats were anaesthetized with 4% halothane and road tunnel dust (RTD, solved in saline at concentrations of 0.15, 0.5, 1.5 or 5 mg/ml) was instilled in a volume of 2 ml/kg body weight to the final concentrations of 0.3, 1, 3, and 10 mg/kg body weight 21. In addition, rats were instilled with either saline solution (0.9% NaCl), or urban dust preparation EHC-93 (urban dust PM sample collected from Environmental Health Centre in 1993, Ottawa, Canada) at a final concentration of 10 mg per kg body weight through a cannula inserted into the trachea just above the bifurcation. Road tunnel dust (RTD) and urban air PM sample (Ottawa dust; EHC-93) were chemically characterized and described previously 21. Briefly, RTD is an integrated PM sample, consisting of coarse and fine fractions, collected at the exit of a motorway tunnel at Hendrik-Ido-Ambacht (HIA). This was collected on polyurethane foam (PUF) using a high volume cascade impactor (HVCI) 22. EHC-93, an urban air PM sample recovered by vacuuming of bag-house filters of the Environmental Health Centre in Ottawa in Canada was used as a second ambient PM sample. The chemical composition and biological reactivity of EHC-93 have been described earlier 2324.

Experiments were approved by the Animal Ethics Committee (IUCAC) of the Dutch National Vaccine Institute (NVI), Bilthoven, The Netherlands.

Tissues were freeze dried for three days. The dried tissues were pulverized using a small mortar and divided into two portions for either mRNA isolation or protein analysis. Fractions for protein analyses were dissolved in 50 mM n-octyl β-D-glucopyranoside (Sigma-Aldrich) in HN-Buffer (25 mM HEPES, 175 mM NaCl, pH 7.7), vortexed, and centrifuged twice (10 min, 13000 rpm). Total protein content of the tissue homogenates was spectrophotometrically determined using the Biorad DC Protein Assay system according to the manufacturer's instructions (Bio-Rad Laboratories B.V., Veenendaal, The Netherlands).

Plasma levels of thrombin-antithrombin complexes (TAT) were measured by a specific murine sandwich TAT ELISA as described previously 20.

Total RNA was isolated using a Trizol method according to the single-step method previously described 25, with minor modifications (Sigma-Aldrich). Five to ten mg freeze-dried tissue was dissolved in 1 mL Tri Reagent. Samples were stored at -80° for precipitation steps, RNA was washed with 80 %(v/v) ethanol and concentrations were spectrophotometrically measured in RNase free water. cDNA was synthesised using the Avian Enhanced First Strand Synthesis kit (Sigma-Aldrich) according to the manufacturer's instructions. mRNA levels for TF and TM were measured on a Light cycler system 1.2 (Roche, Woerden, The Netherlands) using the SYBR premix Ex Taq kit (Takara Bio Inc., Shiga, Japan) according to the manufacturer's instructions and using the following primers: TM forward 5'-GTCACGGTCTCGACAG, TM reverse 5'-GCAGCGTTTGAAAGTCC, TF forward 5'-GAAGAACACCCCGTCG, and TF reverse 5'-GTTCGTCCTAACGTGACA. Quantification was done relative to the household gene glyceraldehyde 3 phosphate dehydrogenase (GAPDH) using forward 5'-TCCCAGAGCTGAACGG and reverse 5'-GAAGTCGCAGGAGACA primers.

Levels of TF and TM are expressed as ratios of TF/GAPDH or TM/GAPDH and each data point represents an average of 3 measurements.

TF activities in tissue homogenates were determined using a home-made activity assay as previously described 26. In brief, dissolved tissue homogenates with a concentration of 1 mg/mL total protein were diluted 20 times (brain, lung and aorta) or 5 times (all other tissues) in HN-buffer. A reference curve was prepared from Innovin (Dade Behring Holding GmbH, Liederbach, Germany), starting with 5 pM and diluted serially 7 times, also in HN buffer. Samples were incubated for 45 minutes at 37°C in the presence of recombinant factor VII (FVII) (Novo Nordisk, Bagsværd, Denmark), 0.2 mM 20/80 PS/PC vesicles, 1 U/mL Bovine factor X (Sigma-Aldrich) and 100 mM Ca²⁺. The formation of factor Xa was then measured kinetically using the chromogenic substrate S 2765 (Chromogenix, final concentration of 0.7 mg/mL diluted in 50 mM Tris-HCl, 175 nM NaCl, 30 mM Na₂EDTA, pH 7.4) by measuring the OD at 405 nm each 15 seconds, for 15 minutes at 37°C.

The Calibrated Automated Thrombogram (CAT, Thrombinoscope, the Netherlands) was used to determine the contribution of mouse and rat tissue homogenates to thrombin generation. We adapted the protocol from the recording of thrombin generation curves in platelet poor plasma as described previously 27: 15 μl of tissue homogenate (5 mg/mL total protein) was added to 80 μL of platelet poor pooled human plasma (normal pool plasma, NPP) (University Hospital Maastricht), which consisted of plasma from 80 healthy volunteers. Thrombin generation was either triggered by adding 5 pM TF, 4 μM phospholipids (PL) (TF/PL) and 16.7 mM Ca⁺² to the reaction mixture, or without addition of TF/PL. In the absence of additional TF/PL triggering of thrombin generation depends entirely on procoagulant molecules, such as TF, present in the tissue homogenate. Furthermore, activation of the coagulation cascade via the contact system will occur in case no additional TF/PL are added and TF is not present in the tissue homogenate added. The following five parameters were derived from a thrombin generation curve: lag time, time to peak, peak height, and endogenous thrombin potential (ETP, the area under the curve). The lag time is merely determined by TF and factor VII, whereas both peak height and ETP are dependent on fibrinogen, prothrombin, and antithrombin (Dielis, Spronk et al. unpublished data). Both mouse and human active site inhibited FVIIa were a kind gift of Dr. Petersen (Novo Nordisk A/S, Denmark).

Data are presented as mean with standard deviation. Differences between groups were assessed using non-paired Student's t-test for normal distribution or Mann-Whitney U test when distribution was not normal. P values < 0.05 were considered statistically significant. Statistical analyses were performed using SPSS version 12.01 for Microsoft Windows.

To assess the overall procoagulant activity of various organs, tissue homogenates were analyzed using an in vitro plasma-based fluorogenic thrombin generation assay. This method uses TF (5 pM), phospholipids (PL, 4 μM) (the so called PPP-reagent), and CaCl₂ (16.7 mM) to trigger thrombin generation in platelet-poor human plasma. From the resulting thrombin generation curve several parameters can be derived, including the lag time (defined as the time to reach 1/6 of the maximum thrombin formed), the peak height (the maximum amount of thrombin formed), and the area under the curve, also known as the endogenous thrombin potential (ETP). For normal human pool plasma a lag time of thrombin generation of 2.56 ± 0.09 min and an ETP of 1592 ± 67 nM.min were observed.

Modification of this method consisted of the addition of normal mice tissue homogenate to platelet-poor human plasma, along with the additional reagent (5 pM TF, 4 μM, 16.7 nM CaCl₂) to trigger thrombin generation. By doing so, addition of lung homogenate (ETP: 194 ± 159 nM.min) caused a 8-fold reduction in thrombin generation compared to platelet-poor human plasma alone (ETP: 1591 ± 67 nM.min, p < 0.01, Figure 1, Panel A). Furthermore, compared to platelet-poor human plasma alone heart homogenate (ETP: 874 ± 75 nM.min, p < 0.01) attenuated thrombin generation, whereas the addition of brain (ETP: 1802 ± 43 nM.min), kidney (ETP: 1631 ± 81 nM.min), liver (ETP: 1631 ± 49 nM.min), or spleen (ETP: 1298 ± 202 nM.min) had no significant influence (Figure 1F, white bars).

As stated before, thrombin generation dependents on the addition of TF, PL, and Ca²⁺, although in the presence of only PL and Ca²⁺, thrombin generation can occur as a result of contact activation, which is characterized by a long lag time before thrombin generation (>10 minutes). If, however, tissue homogenates were added without additional TF, PL, and Ca²⁺, thrombin generation in platelet-poor human plasma became dependent on the presence of TF and PL within the added tissue homogenate. Upon addition of mouse tissue homogenates to plasma, thrombin generation in the absence of additional TF, PL, and Ca²⁺, showed considerable variation in lag times and ETPs compared to analysis in the presence of TF, PL, and Ca²⁺ (Figure 1B, E, and 1F). The short lag times observed for the thrombin generation curves after addition of lung, brain, and kidney were comparable between analysis in the presence or absence of additional TF, PL, and Ca²⁺ (Figure 1E), suggesting that sufficient amounts of TF were present in these tissues to trigger thrombin generation. These findings are compatible with the notion that the constitutive concentrations of TF are high in brain and lung, which is in agreement with the observed TF activity in the chromogenic assay (Figure 2). In contrast, liver (Figure 1B) and heart produced markedly prolonged lag times in the absence of additional TF, PL, and Ca²⁺, as compared to lung, brain, and kidney (Figure 1E), indicating low TF-concentrations in these homogenates and suggesting that contact activation triggered thrombin generation. This was supported by the observation that no thrombin generation was observed in the presence of the coagulation factor XIIa inhibitor corn trypsin inhibitor (CTI), or when FXII-deficient plasma was used instead of normal plasma (data not shown). The strongest overall procoagulant effect, expressed as ETP, was observed for liver, brain, and kidney, while lung and heart hardly induced any thrombin generation (Figure 1D). Kidney, liver, and spleen homogenates also initiated thrombin generation, as indicated by the ETP-values (Figure 1D), despite the rather low levels of endogenous TF (Figure 2).

To ensure that murine TF activity was indeed an important determinant of thrombin generation in platelet-poor human plasma, a separate experiment was performed in which active site inhibited factor VIIa (ASIS) was added to plasma containing mouse lung homogenate. In all cases, both human and murine ASIS prolonged the lag time dose dependently, impairing thrombin generation completely at the highest dose of ASIS (data not shown), demonstrating that TF from mouse tissue homogenates triggers thrombin generation in platelet-poor human plasma and that species specificity is not a limitation in this new method.

Overall, despite the presence of relatively large amounts of endogenous TF (Figure 2: TF activity in lung homogenate and Figure 1B lag times from thrombin generation analysis in the presence of lung homogenate) thrombin generation with addition of lung homogenate was reduced for both the analysis with and without addition of TF, PL, and Ca²⁺. These results suggested the presence of one or more inhibitors in lung and heart tissue with potential factors being glycosaminoglycans, components of the protein C pathway, or tissue factor pathway inhibitor (TFPI). While the latter cannot be ruled out with certainty, heparinase treatment of lung homogenates did not influence the inhibitory effect of lung tissue on thrombin generation (data not shown), suggesting no influence of endogenous glycosaminoglycans. Measuring thrombin generation in protein C-deficient human plasma, however, resulted in increased thrombin generation (e.g. higher ETP values compared to normal human pool plasma) after addition of lung homogenate (ETP: 1818 ± 176 nM.min), indicating that the observed reduction in thrombin generation after addition of lung homogenate to plasma was most likely dependent on a component of the protein C pathway (Figure 1C and 1F grey bars). The same was true for the addition of heart homogenate (ETP: 1619 ± 35 nM.min) (Figure 1F).

Since TM is one of the key factors in protein C activation we next analyzed organ homogenates from TM^Pro/Pro animals which carry a single amino acid functional mutation in the thrombin binding domain that markedly diminishes the protein C cofactor activity of TM 19. Thrombin generation in normal pool plasma with addition of TM^pro/pro organs was comparable to levels obtained for human plasma alone (lung ETP: 1686 ± 209 nM.min, heart ETP: 1767 ± 200 nM.min, liver ETP: 2017 ± 66 nM.min), as well as wild-type organs added to protein C-deficient plasma (Figure 1D), indicating the presence of active TM in lung and heart homogenates. On the basis of these data it is likely that the inhibitory effect of specific organs in the thrombin generation assay is due to functionally active TM.

To validate the new functional method for TF and TM activity in murine tissues a LPS-induced endotoxemia mouse model was used. One of the key characteristics of LPS-induced endotoxemia is the activation of coagukation as well as increased TF expression in lung tissue. Activation of coagulation upon LPS-induced endotoxemia was confirmed by measuring thrombin-antithrombin (TAT) complex levels 6 hours after lipopolysaccharide (LPS)-treatment of C57Bl/6 mice. LPS-stimulated mice showed a 20-fold increase in plasma TAT-levels (204.8 ± 123.3 ng/mL), compared to non-treated control animals (9.2 ± 7.7 ng/mL, p < 0.0001).

To determine the procoagulant activity of LPS challenge at the organ level we compared TF activities in different organs from endotoxemic and control mice (Figure 2): TF activity in lungs of endotoxemic animals (25.8 ± 5.3 pM) was significantly higher than in lungs of control mice (12.7 ± 1.7 pM, p < 0.05), whereas the activity in other organs was comparable between both groups (Figure 2). In accordance, TF mRNA levels in lung tissue were significantly increased after endotoxemia (1.8 ± 0.1 vs. 1.4 ± 0.3, p < 0.02; Figure 3A), whereas the levels in other organs such as the heart (0.07 ± 0.03, Figure 3B) remained at a level comparable to controls (0.10 ± 0.03, p > 0.1) as expected.

Comparing thrombin generation curves recorded in either the presence or the absence of TF, PL, and Ca²⁺, demonstrated a significant induction of thrombin generation for lung homogenates from mice subjected to LPS challenge (ETP: +TF: 940 ± 523, -TF: 692 ± 369 nM.min, Figure 4A), compared to controls (ETP: +TF:194 ± 159 nM.min, -TF: 103 ± 83 nM.min, p < 0.01, Figure 1F). These results are suggestive of attenuated inhibitory potential (e.g. decreased TM levels) or increased TF activity in lung tissue due to endotoxemia. The opposite influence of LPS was observed for heart tissue, which in the absence of PPP-reagent showed a significantly prolonged lag time after LPS challenge (15 ± 4 min, Figure 4C) compared to control heart tissue (11 ± 1 min, p < 0.05, Figure 4C). This later observation suggest decreased TF expression in the heart upon LPS induced inflammation, as observed previously by Luther et al. 28.

No changes were observed for brain (ETP: +TF 1795 ± 49 nM.min, -TF: 1488 ± 57 nM.min) and heart (ETP: +TF: 785 ± 90 nM.min, -TF: 179 ± 58 nM.min) (Figure 4D) from LPS-treated animals compared to control tissues (Brain ETP: +TF 1802 ± 43 nM.min, -TF: 1434 ± 60 nM.min; Heart ETP: +TF: 874 ± 74 nM.min, -TF: 144 ± 31 nM.min)(Figure 1F). Note that analyses were done both in the presence (+TF) and absence (-TF) of additional TF.

The increased thrombin generating activity of lung tissue after LPS treatment is suggestive for reduced availability of TM in the tissue homogenate, based on the aforementioned dependence of thrombin generation inhibition by TM in tissue; this conclusion is supported by the observed reduced TM mRNA expression in lung after LPS challenge (0.16 vs. 0.03, p < 0.01, Figure 3C). In contrast, we could not detect TM mRNA in other tissues.

Recent data showed enhanced expression of TF mRNA in lungs of spontaneously hypertensive rats after exposure to PM 29. Here, we analyzed the pro- and anticoagulant activity in rat lung tissue at different time points (4 and 48 hours) after exposure to two types of PM: road tunnel dust (RTD) and urban dust (EHC-93). At 4 and 48 hours after exposure to RTD, TF activities in lungs showed a trend to increase with increasing concentrations of RTD. TF activity was significantly increased at the highest concentration of PM (10 mg RTD per kg BW) compared to saline instillation (at 4 hours: 1045 ± 472 pM vs. 397 ± 231 pM) (Figure 5). No differences in TF activity were observed between saline and EHC-93 (10 mg/kg) instillation at 4 hours after exposure (587 ± 260 pM), whereas TF activity was markedly increased (1870 ± 693 pM, p < 0.01) at 48 hours after exposure.

Thrombin generation (Table 1) was increased with rising concentrations of PM, with the difference between the dose of 10 mg/kg BW and saline control being significant at 4 hours (RTD: ETP 390 ± 148 nM.min, EHC-93: ETP 264 ± 64 nM.min vs. saline: ETP 186 ± 80 nM.min, p < 0.01) and 48 hours (RTD: ETP 1441 ± 289 nM.min, EHC-93: ETP 1719 ± 103 nM.min vs. saline: ETP 174 ± 61 nM.min, p < 0.0001). The increase in thrombin generation at 48 hours compared to 4 hours (with almost equal TF activities and lag times) strongly suggests a loss in TM activity at the 48 hrs time point.