Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters
http://circ.ahajournals.org//subscriptions/Subscriptions: Information about subscribing to Circulation is online at
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
Early Atherosclerosis Exhibits an EnhancedProcoagulant State
Julian Ilcheff Borissoff, MD; Sylvia Heeneman, PhD; Evren Kilinç, MSc; Peter Kassak, MSc;Rene Van Oerle, BSc; Kristien Winckers, MD; Jose W.P. Govers-Riemslag, PhD;
Karly Hamulyak, MD, PhD; Tilman M. Hackeng, PhD; Mat J.A.P. Daemen, MD, PhD;Hugo ten Cate, MD, PhD; Henri M.H. Spronk, PhD
Background—Thrombin generation in vivo may be important in regulating atherosclerotic progression. In the presentstudy, we examined for the first time the activity and presence of relevant coagulation proteins in relation to theprogression of atherosclerosis.
Methods and Results—Both early and stable advanced atherosclerotic lesions were collected pairwise from each individual(n�27) during autopsy. Tissue homogenates were prepared from both total plaques and isolated plaque layers, in whichthe activity of factors (F) II, X, and XII and tissue factor was determined. Microarray analysis was implemented toelucidate local messenger RNA synthesis of coagulation proteins. Part of each specimen was paraffin embedded, andhistological sections were immunohistochemically stained for multiple coagulation markers with the use of commercialantibodies. Data are expressed as median (interquartile range [IQR]). Tissue factor, FII, FX, and FXII activities weresignificantly higher in early atherosclerotic lesions than in stable advanced atherosclerotic lesions. Endogenous thrombinpotential and thrombin-antithrombin complex values consolidated a procoagulant profile of early atherosclerotic lesions(endogenous thrombin potential, 1240 nmol/L � min [IQR, 1173 to 1311]; thrombin-antithrombin complex, 1045 ng/mg[IQR, 842.6 to 1376]) versus stable advanced atherosclerotic lesions (endogenous thrombin potential, 782 nmol/L � min[IQR, 0 to 1151]; thrombin-antithrombin complex, 718.4 ng/mg [IQR, 508.6 to 1151]). Tissue factor, FVII, and FXcolocalized with macrophages and smooth muscle cells. In addition, multiple procoagulant and anticoagulant proteaseswere immunohistochemically mapped to various locations throughout the atherosclerotic vessel wall in both early andadvanced atherosclerotic stages.
Conclusions—This study shows an enhanced procoagulant state of early-stage atherosclerotic plaques compared withadvanced-stage plaques, which may provide novel insights into the role of coagulation during atherosclerotic plaqueprogression. (Circulation. 2010;122:821-830.)
Atherosclerosis is widely recognized as a chronic inflam-matory disease.1 Rupture of an atherosclerotic plaque is
considered the predominant underlying cause of acute athero-thrombotic events such as myocardial infarction, ischemicstroke, and vascular death. A close relation between bloodcoagulation and atherosclerosis2,3 is supported by studiesrevealing the presence of specific coagulation proteinswithin an atherosclerotic lesion. Tissue factor (TF) andfactor (F) VII, of which the complex is the principalinitiator of coagulation in vivo, are expressed on macro-phages and vascular smooth muscle cells (SMC) within thearterial vessel wall and atherosclerotic lesion.4,5 Bothproteins potentially participate in multiple proatherogenic
processes such as migration and proliferation of SMC,6
inflammation, and angiogenesis.7 In addition to the singleeffects of each protein, the local interaction betweenmacrophage/SMC-derived TF and FVII may provide acatalytic complex for subsequent generation of thrombin andfibrin, of which the latter is also detectable in atheroscle-rotic lesions.8,9 The procoagulant condition of the athero-sclerotic lesion may be further enhanced by the presence ofvarious proinflammatory cytokines (eg, tumor necrosisfactor-�, interleukin-110), which may downregulate localexpression of anticoagulant proteins such as thrombo-modulin and the endothelial protein C receptor on endo-thelial cells.11
Received September 4, 2009; accepted June 28, 2010.From the Laboratory for Clinical Thrombosis and Hemostasis, Department of Internal Medicine (J.I.B., E.K., P.K., R.V.O., K.W., J.W.P.G.-R., H.t.C.,
H.M.H.S.); Department of Pathology (S.H., M.J.A.P.D.); Department of Biochemistry (K.W., T.M.H.); and Division of Hematology, Department ofInternal Medicine (K.H.), Cardiovascular Research Institute Maastricht, Maastricht University Medical Center, Maastricht, the Netherlands.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.907121/DC1.Correspondence to Julian Ilcheff Borissoff, MD, Laboratory for Clinical Thrombosis and Hemostasis, Cardiovascular Research Institute Maastricht, Maastricht
Clinical Perspective on p 830Thrombin, a key enzyme in blood coagulation, may also
play a critical role in many processes related to the develop-ment, progression, and atherothrombotic potential of athero-sclerotic plaques.12 Direct evidence for the role of thrombin inthe atherogenic process comes from experiments showingreduced progression of atherosclerosis in apolipoprotein E�/�
mice on pharmacological inhibition of thrombin.13 Moreover,decreased expression of TF pathway inhibitor (TFPI) on anapolipoprotein E�/� background increased the atheroscleroticburden.14
Because of the reported involvement of procoagulant andanticoagulant coagulation factors during plaque progression,we hypothesized that the overall distribution and activity ofcoagulation proteins in the arterial vessel wall correlate withthe extent and progression of atherosclerotic lesions. More-over, we hypothesized that the amount of thrombin that canbe generated from atherosclerotic tissue homogenates de-pends not only on the amount of TF but also on the presenceand activity of other coagulation proteins that either amplifyor dampen thrombin generation. Hence, we studied thelocalization of all coagulation proteins, in addition to theTF/FVII complex, on histologically defined early and stableadvanced atherosclerotic lesions. In addition to thrombingeneration, we determined the procoagulant activity of sev-eral coagulation proteins in the same lesions.
MethodsPatient Characteristics and Tissue SpecimensThe tissue specimens were obtained from the Maastricht PathologyTissue Collection. Collection, storage, and use of tissue and patientdata were performed in agreement with the Code for ProperSecondary Use of Human Tissue in the Netherlands (http://www.fmwv.nl). Both early atherosclerotic lesions (EAL) and stableadvanced atherosclerotic lesions (SAAL) were collected pairwisefrom each corresponding individual (n�27) during postmortemdissection of the abdominal aortas within 8 hours of death. Autopsyspecimens were obtained from adult men and women with an agerange of 45 to 84 years (mean, 55 years). Clinical characteristics ofthe patients are provided in Table 1. The cause of death was diverse(eg, myocardial infarction, stroke). Individuals with sepsis or cancerwere excluded. All tissue specimens were histologically evaluated onhematoxylin and eosin–stained sections (4 �m). Plaque subtypeswere determined in compliance with the modified American HeartAssociation classification, based on morphological description, pro-posed by Virmani et al.15 Because one of the main goals in this studywas to discriminate the overall prothrombotic potential of athero-sclerotic lesions between early and advanced stages of development,we classified the plaques as follows: intimal thickenings and xan-thomas are uniformly termed EAL, whereas all types of stableadvanced plaques are termed SAAL. Complicated lesions, includinglesions with intraplaque hemorrhage, a surface defect, and/or throm-botic deposit, were not included in this study.
Preparation of Tissue HomogenatesA section of each of the collected specimens (27 EAL/27 SAAL,obtained in pairs from n�27) was snap-frozen on collection. Snap-frozen atherosclerotic tissues were freeze-dried for 3 days andpulverized, and subsequently the tissue powders were dissolved in50 mmol/L N-octyl �-D-glucopyranoside (Sigma-Aldrich) in HNbuffer (25 mmol/L HEPES, 175 mmol/L NaCl, pH 7.7), vortexed,
and centrifuged twice (10 minutes, 13 000 rpm). Total proteincontent of the tissue homogenates was spectrophotometrically deter-mined with the use of the Biorad DC Protein Assay system accordingto the manufacturer’s instructions (Bio-Rad Laboratories B.V.,Veenendaal, Netherlands). All samples were further diluted into afinal concentration of 5 mg/mL.
Effect of Time Delay Between Death andPostmortem Examination on CoagulationProtein ActivitySee Methods in the online-only Data Supplement.
EAL and SAAL Layer Preparationand HomogenizationSee Methods in the online-only Data Supplement.
Thrombin Generation, Prothrombin, FX, andFXII Activity Assays, Thrombin-AntithrombinComplex Levels, TF Activity Assay, and TFPIAntigen AssayThe calibrated automated thrombogram (Thrombinoscope, the Neth-erlands) was used to determine the contribution of atherosclerotictissue homogenates to thrombin generation in human plasma (intriplicate; interassay coefficient of variation �10%). For additionalinformation, see Methods in the online-only Data Supplement.
Effect of Phospholipid Concentration on ThrombinGeneration in Normal Arterial VesselWall HomogenatesSee Methods in the online-only Data Supplement.
RNA Isolation and Quantification, MicroarrayHybridization, and Data AnalysisSee Methods in the online-only Data Supplement.
Immunohistochemical and ImmunofluorescenceStainings and Immunohistological EvaluationSee Methods in the online-only Data Supplement.
Statistical AnalysisData analysis was computed with SPSS, version 17.02 (SPSS Inc,Chicago, Ill) and Prism, version 5.00 (GraphPad Software Inc, SanDiego, Calif). Results are expressed as median (interquartile range[IQR]). An exact-distribution Wilcoxon 2-sample test was used forall intraindividual comparisons. A 2-tailed P�0.05 was considered
Table 1. Patient Clinical Characteristics of Autopsy Cases FromWhich Paired EAL and SAAL Were Obtained and Examined
Clinical Characteristics All (n�27)
Age, y 55.4�9.2
Male sex 16 (59.3)
Hypertension 14 (51.9)
Smoking 11 (40.7)
Hypercholesterolemia 10 (37.0)
Diabetes mellitus 3 (11.1)
Family history 17 (63)
Previous myocardial infarction 8 (29.6)
Values are n (%) except age.
822 Circulation August 24, 2010
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
statistically significant. Repeated-measures ANOVA was used toassess differences in coagulation protein activity over time.
ResultsEAL Exhibit Higher Functional Activity of KeyCoagulation Proteins Than SAAL Ex VivoWe determined the dependence of the thrombin-generatingpotential of both EAL and SAAL on their prothrombin, FX,and FXII content. From all 54 specimens (27 pairs, EAL andcorresponding SAAL) that we examined, prothrombin activ-ity was detected in only 11 samples. From the latter 11samples with detected activity, EAL specimens had signifi-cantly higher prothrombin activity at 0.0% (IQR, 0.0 to7.761) compared with their paired SAAL at 0.0% (IQR, 0.0 to0.0) (Figure 1A; Wilcoxon 2-sample test, 2-tailed exactP�0.05). The activity of FX revealed a similar trend, with asignificant 3-fold upregulation in EAL at 0.276% (IQR, 0.164to 0.536) compared with SAAL at 0.136% (IQR, 0.054 to0.237) (Figure 1B; Wilcoxon 2-sample test, 2-tailed exactP�0.05). Furthermore, FXII also demonstrated significantlyhigher activity levels in EAL of 2.636% (IQR, 1.344 to3.372) compared with levels in SAAL of 0.930% (IQR, 0.337to 1.526) (Figure 1C; Wilcoxon 2-sample test, 2-tailed exactP�0.05).
EAL Demonstrate a 3-Fold Increase in TF ActivityVersus SAALTo better appreciate the procoagulant potential of these 2 setsof atherosclerotic plaque homogenates, we assessed theactivity of TF, which is known to be a pivotal trigger ofcoagulation in vivo. TF activity was �3-fold higher (0.036pmol/mg [IQR, 0.017 to 0.055]) in EAL compared withSAAL (0.009 pmol/mg [IQR, 0.005 to 0.022]) (Figure 1D;Wilcoxon 2-sample test, 2-tailed exact P�0.05). Twenty-sixof 27 EAL homogenates indicated elevated TF activity levelscompared with their corresponding SAAL specimens.
Notably, the 8-hour window between death and postmor-tem collection did not significantly affect the activity of TF,FII, FX, and FXII in atherosclerotic lesions harvested atvarious time points: 0 (baseline), 2, 4, and 8 hours. Nosignificant differences were found between the different timepoints of all tested proteins and also compared with baselinevalues (Figure I in the online-only Data Supplement),strongly suggesting that the postmortem values reflectedactual coagulation activity in vivo.
Shift of the TF/TFPI Ratio Suggests an IncreasedAtherothrombotic Tendency in EALTFPI is a potent natural inhibitor of the TF-driven pathway ofthe coagulation cascade and also plays an important role inregulating inflammation. Furthermore, it has been shown thatTFPI modulates thrombus formation in experimental modelsin vivo,16 primarily by attenuating the procoagulant activityand overexpression of TF.17,18 Therefore, we tested the levelsof TFPI by utilizing a homemade enzyme-linked immunosor-bent assay. An �1.6-fold significant increase in TFPI antigenlevels was found in EAL compared with SAAL. EALdemonstrated TFPI activity equal to 0.089 nmol/L per milli-gram (IQR, 0.072 to 0.140), whereas SAAL showed 0.056nmol/L per milligram (IQR, 0.030 to 0.088) (Figure 1E;Wilcoxon 2-sample test, 2-tailed exact P�0.05). Despite thehigher levels of TFPI antigen in the EAL homogenates, inSAAL the TF/TFPI balance in the early lesions remained infavor of TF, shown by the higher TF/TFPI ratios in EALhomogenates (Figure 1F; Wilcoxon 2-sample test, 2-tailedexact P�0.05).
Enhanced Thrombin Generation in EALIn the absence of TF and entirely dependent on the proco-agulant molecular content in the tissue homogenate, all 27EAL induced thrombin formation in normal pooled plasma,showing significantly higher values (1240 nmol/L � min
Figure 1. Activity of coagulation pro-teases in paired EAL and SAAL homoge-nates. A, FII activity in FII-deficientplasma, assessed via modified calibratedautomated thrombogram measurement,indicates significantly higher levels in EAL.B, Upregulated FX activity in EAL homog-enates compared with the correspondingSAAL. FX activity was significantly ele-vated (�3-fold increase) in EAL homoge-nates compared with SAAL homogenates;Wilcoxon 2-sample test, 2-tailed exactP�0.05. C, FXII activity comparisonbetween EAL and SAAL, showing signifi-cantly higher levels in EAL (�2-foldincrease); Wilcoxon 2-sample test,2-tailed exact P�0.05. D, TF activity is�3-fold higher in EAL homogenates thanin SAAL homogenates; Wilcoxon2-sample test, 2-tailed exact P�0.05. E,TFPI activity shows 1.6-fold increase inEAL. F, TF/TFPI ratio in EAL (0.294 [IQR,0.109 to 0.770]) and SAAL (0.174 [IQR,0.117 to 0.257]); Wilcoxon 2-sample test,2-tailed exact P�0.05.
Borissoff et al Coagulation Markers in Atherosclerosis 823
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
[IQR, 1173 to 1311]) compared with SAAL (782 nmol/L � min [IQR, 0 to 1151]) (Figure 2B; Wilcoxon 2-sampletest, 2-tailed exact P�0.05). Twenty-six EAL induced higherendogenous thrombin potential than their corresponding ad-vanced atheromas. For the SAAL, 10 lesions did not triggerany thrombin generation.
Furthermore, EAL showed a significantly increased throm-bin generation potential compared with areas of normal aortaobtained from the same individuals (263.3 nmol/L � min[IQR, 117.8 to 350.3]; Wilcoxon 2-sample test, 2-tailed exactP�0.0001) (Figure 2B). In addition, SAAL also demon-strated significantly higher endogenous thrombin potentialthan their paired normal vessel homogenates (Wilcoxon2-sample test, 2-tailed exact P�0.0053), thus consolidatingthe procoagulant tendency in early atherosclerosis.
Thrombin-Antithrombin Complex LevelsAdditionally Point to Higher ThrombinGeneration in EAL HomogenatesOnce generated, thrombin is inhibited on binding to anti-thrombin, thus forming a stable thrombin-antithrombin(TAT) complex. TAT complexes are considered a marker ofin vivo intravascular thrombin generation; therefore, the maingoal of this experiment was to assess whether there was anexcess of activated FII generation in EAL in situ comparedwith their matched SAAL. The concentration of TAT com-plexes in EAL was significantly higher (1045 ng/mg [IQR,842.6 to 1376]) compared with their paired SAAL homoge-nates (718.4 ng/mg [IQR, 508.6 to 1151]) (Figure 2C;Wilcoxon 2-sample test, 2-tailed exact P�0.05), confirming amore procoagulant state in EAL.
Layer-Selective Analysis of Coagulation FactorActivities Consolidated a More Procoagulant Stateof EAL Versus SAALTo provide better insight into the procoagulant properties ofthe atherosclerotic lesions, we undertook a more selective,
layer-specific analysis in which the potential procoagulanteffects of the different vessel wall layers were studied. Theactivity of coagulation factors was analyzed in tissue homog-enates prepared from tunica intima, media, and adventitia(histologically controlled anatomic separation; Figure 3A).All 3 layers were harvested in 42 specimens (21 pairs of EALand SAAL from the original tissue collection). Endogenousthrombin potential values in all layers of EAL were found tobe significantly higher (intima: 1489 nmol/L � min [IQR 1353to 1680]; media: 1734 nmol/L � min [IQR, 1256 to 1983];adventitia: 1872 nmol/L � min [IQR, 1655 to 2171]) com-pared with the corresponding SAAL layers (intima: 437.9nmol/L � min [IQR, 290.3 to 549.9]; media: 392.1 nmol/L � min [IQR, 219.7 to 680.9]; adventitia: 524.1 nmol/L � min[IQR, 394.1 to 787.7]) (Figure 3B; Wilcoxon 2-sample test,2-tailed exact P�0.05, all). This strongly pronounced proco-agulant state of the EAL layers was additionally confirmed bysignificantly elevated prothrombin, FX, and FXII levels(Figure 3C to 3E). Intimal layers of both EAL and SAALshowed comparable TF activity, whereas TF was signifi-cantly increased in media and adventitia of EAL versusSAAL (Figure 3F). Although they demonstrated comparableactivities in terms of TF, EAL intimal layers containedsignificantly higher TFPI levels (Figure 3G), yielding asignificantly lower TF/TFPI ratio in EAL compared withSAAL. EAL and SAAL media layers did not significantlydiffer in TFPI levels, whereas TFPI in EAL adventitia wassignificantly higher compared with SAAL (Figure 3G). Tu-nica adventitia exhibited the most procoagulant phenotype ofall vessel wall layers in terms of thrombin generation. Itsvalues in both EAL and SAAL were significantly higher thanthose measured in tunica intima and media.
Gene Expression of Coagulation Genes in EALVersus SAALTo better explore to what extent and which coagulationproteins are expressed on the genome level within the arterial
Figure 2. Effect of phospholipid concentration on thrombin generation. Overall procoagulant activity of atherosclerotic plaque homoge-nates and normal vessels is shown, assessed by means of thrombin generation and/or TAT complexes. A, Influence of increasingphospholipid concentrations (1, 2, 3, 4, 5, 10, 20, 30, 40, and 50 �mol/L) on thrombin generation in normal pool plasma triggered by 1pM TF (without addition of plaque homogenates). At phospholipid concentrations of �4 �mol/L, thrombin generation is independentfrom additional phospholipids present in the measured sample (such as is present in the added plaque homogenate) (P value was cal-culated by repeated-measures ANOVA). Furthermore, thrombin generation in EAL and/or SAAL homogenates was established at a finalphospholipid concentration of 4 �mol/L, irrespective of the plaque type studied. ETP indicates endogenous thrombin potential. B,Thrombin generation assessed in paired EAL and SAAL and areas of normal aorta. EAL and SAAL show significantly higher thrombingeneration levels compared with normal arterial vessels. These data consolidate the procoagulant state in early atherosclerosis but fur-thermore indicate that the prothrombotic tendency in EAL is not dependent on variations in the cellular density/phospholipid content asa result of vessel wall structure alterations that occur on atherosclerotic progression. HAV indicates healthy arterial vessels. C, TATcomplex levels measured in paired EAL and SAAL homogenates.
824 Circulation August 24, 2010
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
vessel wall, gene expression profiles of EAL and SAAL wereobtained with the use of microarray analysis. In a separate setof patients, early and advanced carotid lesions were collectedfrom the same patient (at autopsy), and fold changes in geneexpression were assessed by comparing the advanced lesionswith the early lesions. The results indicated that severalcoagulation factor genes were expressed in both types ofatherosclerotic lesions. After correction for multiple testing,14 coagulation genes showed significant differential tran-script levels between EAL and SAAL. Figure 4 demonstratesthe relative mRNA levels, described as the SAAL/EAL ratio.Of these 14 differentially regulated genes, 6 were upregulatedin EAL (expressed as fold change ��1.0), whereas 8 wereupregulated in SAAL (expressed as fold change ��1.0)(Table I in the online-only Data Supplement). Fold changesranged from �1.13 to �2.96 for the upregulated genes inEAL and from 1.08 to 1.29 for the upregulated genes inSAAL. Additional information is provided in Table I in theonline-only Data Supplement.
Immunohistochemical Staining: EALIn EAL, moderate (fibrinogen/fibrin, FIX, TFPI) to strongpositivity for von Willebrand factor, FX, prothrombin/throm-bin, protein S, and activated protein C (APC) was observed inthe endothelial luminal cells, indicated by a sharp demarca-tion of the endothelial lining (Figure 5 and Table 2). Inaddition, a positive focal endothelial distribution for TF,FVII, FXII, FXI, kallikrein, and thrombomodulin was shown.Macrophages and foam cells stained intensely positive forTF, FVII, FX, prothrombin/thrombin, kallikrein, and FXI.Despite the fact that other coagulation proteins such as FXII,FIX, protein S, protein C, and APC were also expressed bymacrophages and foam cells, their expression or immunore-activity was either scarce or focal. Furthermore, EAL werecharacterized by TF, FVII, and FX expression throughout theSMC-rich intima. Medial SMC-associated FVII was locatedin the cytoplasm and not on the membrane. FXII and FIIshowed enhanced expression in medial SMC. TF, FVII, FX,fibrin, kallikrein, thrombomodulin, and TFPI were also asso-
Figure 3. Layer-selective coagulation factor activity analysis, presenting the procoagulant state of tunica intima, media, and adventitiain EAL vs SAAL. A, Hematoxylin and eosin–stained sections, demonstrating histologically controlled layer preparation and confirmingthe anatomy of the desired vessel wall layer. The activities of coagulation proteins were then tested in tunica intima, media, and adven-titia, prepared from the harvested layers of paired EAL and SAAL: endogenous thrombin potential (ETP) (B); prothrombin (C); FX (D);FXII (E); TF (F); and TFPI (G). TF/TFPI ratio is demonstrated in H. *Statistical significance (Wilcoxon 2-sample test, 2-tailed exactP�0.05). INT indicates tunica intima; MED, tunica media; and ADV, tunica adventitia.
Borissoff et al Coagulation Markers in Atherosclerosis 825
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
ciated with medial SMC, whereas FIX demonstrated a morepatchy expression. Within the adventitia, the vasa vasorumexterna showed positive staining for most of the studiedfactors, whereas the fibroblasts were positively associatedwith FX, prothrombin/thrombin, kallikrein, von Willebrandfactor, and FXII.
An extensive overview of all single- and double-stainingobservations in EAL and SAAL is provided in Figure 5,Table 2, and Figure II in the online-only Data Supplement.
Immunohistochemical Staining: SAALIn atherosclerotic tissues classified as SAAL, the endothelialluminal lining was moderately positive for FIX, thrombo-modulin, APC, and von Willebrand factor, whereas FXIstained weakly positive (Figure 5). Some endothelial seg-ments demonstrated a focal expression of the anticoagulantprotein protein S. Moreover, all anticoagulant proteins (pro-tein S, thrombomodulin, APC, and TFPI) were found to beassociated with macrophages and foam cells. Furthermore,thrombomodulin, APC, and TFPI were also localized in theendothelial cells of the vasa vasorum and in endothelial cellsof vessels sprouting into the lesions. Besides thrombomodu-lin, intimal and medial SMC contained most of the proco-agulant proteins, as well as thrombin and fibrinogen/fibrin(Figure 5 and Figure II in the online-only Data Supplement).A slight focal association between FIX and XI with intimalSMC was observed. In contrast, SMC of the media stainedmoderately for FXII. Some of the medial SMC stainedpositive for FXI but also showed double positive staining forboth CD68 and anti–smooth muscle actin, suggesting eithertransdifferentiation of SMC into foam-like cells or SMCoutgrowth from mononuclear cells; the latter was reported
Figure 4. Relative mRNA levels of coagulation genes (microar-ray analysis), presented as a SAAL/EAL ratio, indicate the differ-ential expression of various coagulation factor genes in EAL vsSAAL. *Statistical significance (Wilcoxon 2-sample test, 2-tailedexact P�0.05). FVIII-A1 indicates FVIII-associated (intronic tran-script) 1; FVIII, PC, FVIII, procoagulant component; HCII, heparincofactor II; EPCR, endothelial protein C receptor; vWF, von Wil-lebrand factor; and PAI, plasminogen activator inhibitor.
Figure 5. Immunohistochemical stainings for coag-ulation proteases in paired EAL and SAAL; magni-fication �10. Positive staining is presented in red.FII/FIIa indicates prothrombin/thrombin; FGN,fibrinogen; TM, thrombomodulin; PS, protein S;and vWF, von Willebrand factor.
826 Circulation August 24, 2010
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
recently to be a thrombin-promoted action.19 TF and FVII, FX,FXI, and FXII were either weakly or focally present on macro-phages, and FXII was also found on some foam cells. FIXshowed a pronounced focal distribution on both macrophagesand foam cells. Some fibroblasts contained FX, FXII, and fibrin.
None of the SAAL from the current set of lesions showed astrong or even moderately positive staining for TF; this was alsothe case in the necrotic core. The necrotic core revealed a focalpresence for most of the procoagulant proteins, except for thrombin,fibrin, and the anticoagulant APC, which stained weakly positive.
A broad histological evaluation is available in Figure 5,Table 2, and Figure II in the online-only Data Supplement.
Immunofluorescence Staining: Colocalization ofTF/FVII/FX with Macrophages and Vascular SMCDouble immunofluorescence staining with CD68 for macro-phages and anti–smooth muscle actin for SMC suggests thatmost of the macrophages and SMC were involved in the
synthesis of TF, FVII, and FX. The formation of the ternarycomplex TF/FVII/FX is a potent trigger not only for coagu-lation (thrombin) but also for many proinflammatory cell-signaling pathways that are pivotal in cardiovascular disease.Therefore, we also examined the presence of these procoagu-lant proteins on macrophages and SMC by means of immu-nofluorescence staining on corresponding EAL and SAALsections, which revealed that TF/FVII/FX colocalized withboth macrophage/foam cells and SMC, suggesting a localsystem of thrombin generation, which may regulate patho-physiological processes such as cell migration and inflamma-tion. When EAL and SAAL are compared, colocalization isscarcer and more diffuse in SAAL, whereas EAL sectionsshow brighter labeling and denser character (Figure 6).
DiscussionThe present study shows that atherosclerotic plaques exhibitfunctional activity of many coagulation proteins (prothrom-
Figure 6. Immunofluorescence stainingsdemonstrating cellular colocalization ofTF, FVII, and FX with macrophages (M)and VSMC on paired EAL and SAAL;magnification �60. TF, FVII, and FX werestained with fluorescein isothiocyanate(FITC) (green color); macrophages andSMC were stained with rhodamine (Rho)(red color). Blue color shows nuclei. Colo-calizations are demonstrated in yellow/orange color.
Table 2. Presence of Coagulation Proteins in Various Arterial Vessel Wall Compartments Throughout the Early and Advanced Stagesof Atherosclerotic Development
Types and Structures TF FVII FX FII/FIIa Fibrin FXII FXI FIX Kallikrein Thrombomodulin Protein S APC TFPI vWF
EAL
Endothelial cells F F ��� ��� �� F F �� F F ��� ��� �� ���
Macrophages ��� ��� ��� ��� � F ��� F �� � F F � �
The table represents immunohistochemical staining observations in EAL and SAAL, obtained from the same corresponding individuals. FII/FIIa indicatesprothrombin/thrombin; vWF, von Willebrand factor. Staining grading is as follows: �, negative; �, weakly positive; ��, moderately positive; ���, strongly positive;and F, focal (only certain cells stain, not all).
Borissoff et al Coagulation Markers in Atherosclerosis 827
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
bin, FX, FXII, and TF) and represents the first study todemonstrate the presence and distribution of all coagulationproteins in both early and advanced human atheroscleroticplaques. We provide new data pointing to local synthesis ofseveral coagulation proteins within the atherosclerotic vesselwall. Furthermore, we indicate a colocalization of key pro-coagulant proteins with SMC and macrophages, suggestingan active, cell-based coagulation network within the athero-sclerotic plaque. Finally, the principal finding of this study isan enhanced procoagulant profile of EAL compared withSAAL homogenates, consolidated by an elevated thrombingeneration potential and significantly increased TAT com-plex levels in early-stage atherosclerotic tissues. Thus, weprovide novel evidence that may help in widening thethrombogenic spectrum of “high-risk” plaques and suggestthat local coagulation factors may play an important role notonly in contributing to the onset of atherothrombosis but alsoin contributing to progression of the atherosclerotic process.
In contrast to our expectations, these data reject the initialhypothesis that thrombin generation would positively corre-late with progression of atherosclerosis. One possible mech-anism that might explain the abundant presence and func-tional activity of coagulation proteins in the early stage ofatherosclerosis is that many of the coagulation proteins helpto propagate the atheromatous plaque by inducing multipleproatherogenic actions such as cellular adhesion, migration,angiogenesis, and inflammation.7,12 In addition to their pro-thrombotic nature, coagulation proteases induce cell prolifer-ation,6 and the latter are of great importance in determiningthe stability of an atherosclerotic lesion. The abundance ofalmost all (intrinsic and extrinsic) coagulation proteins sug-gests that the generation of thrombin is an active processduring atherogenesis, supporting a major role of thrombin(and possibly fibrin) in this condition. Moreover, the en-hanced procoagulant state of EAL was additionally con-firmed by the layer-selective analysis, which also revealed asignificantly increased prothrombotic phenotype for tunicaintima, media, and adventitia in EAL versus SAAL. Previousreports have also documented that adventitial fibroblasts thatsurround the arterial walls contain high amounts of TF,providing a “hemostatic envelope.”20 Our study confirmedthis latter finding by showing that tunica adventitia was themost procoagulant vessel wall layer of all those tested. Acontributory or distinct effect of other coagulation proteaseson atherogenesis, including FXII, FXI, FIX, and FX, cannotbe ruled out. Published data have associated FXII in bloodwith cardiovascular disease, and although its action hascontributed mainly to promoting arterial thrombosis, itspresence in atherosclerosis and in the vicinity of macrophagesand foam cells suggests cell-directed actions of this protein.Indeed, in vitro work has demonstrated the localization ofcontact system proteins on macrophages,21 suggesting a directeffect on inflammatory pathways. Of interest, although theexpression and activity of FXII diminished on progression ofatherosclerosis in our data, the staining for kallikrein wasmore abundant in the advanced lesions in the vicinity ofmacrophages and foam cells. The latter may be compatiblewith a switch in direction from a procoagulant to a proin-flammatory action of the FXII contact system, as established
in a recent study.22 Switching the action of FXII to aninflammatory direction may explain in part the diminishedthrombin-generating capacity in advanced lesions, which aredominated by inflammatory characteristics (including ele-vated interleukin-6 and tumor necrosis factor-� levels; datanot shown). The apparent loss in thrombomodulin stainingwithin the advanced plaques may also be compatible withincreased inflammatory activity, as proposed previously,11
whereas apparently the vasa vasorum maintains thrombo-modulin in amounts comparable to those in the early lesion.
The abundant presence of coagulation proteins in the earlylesions in particular raises other questions about causes andconsequences. It is well known that the initial stage ofatherosclerotic development (eg, intimal thickening) is char-acterized by enhanced SMC migration and proliferation.23 Incontrast, an advanced stage of atherosclerotic progressionresults in decreased cell density, primarily around the fibrouscaps and necrotic core/lipid pool.24,25 However, the decreasedprocoagulant potential in SAAL was found to be independentof the vessel wall structure alterations that occur on athero-sclerotic progression (Figure 2A and 2B), thus suggesting thatdifferences in protein translocation from circulation towardthe vessel wall, as well as local protein expression bydifferent cell types, may also contribute significantly tovariations in protein levels. Although all coagulation proteinsexcept TF may finally appear in the vessel wall by diffusionfrom the circulating blood, the localization suggests that localsynthesis may be involved. Moreover, the microarray analy-sis clearly shows that multiple coagulation proteins areexpressed on the level of mRNA synthesis in the arterialvessel wall (Figure 4 and Table I in the online-only DataSupplement). However, although some of the coagulationproteins were differentially upregulated in EAL, a similarpicture was also observed for other coagulation genes inSAAL, suggesting that the differences in protein expression(as revealed by immunohistochemistry) and activity levels ofcoagulation proteins were not reflected completely by differ-ences in gene expression levels. Moreover, it is known thatRNA expression profiles do not always correlate with proteinexpression and subsequent biological activity.26 Nevertheless,our data point to local synthesis of several coagulationproteins within the atherosclerotic vessel wall, suggesting thatthis may be part of an active regulatory mechanism, leadingsubsequently to the enhanced procoagulant state in EAL.
The pleiotropic effects of proteases such as thrombin andactivated FX, as well as the cell growth–promoting effectsof fibrin (and its split products), may be also evoked as partof a response to injury mechanism. This response action ofblood coagulation is now well established in inflammatoryconditions like sepsis. As a side effect of this process, theformation of fibrin may serve to protect the early lesions fromrupture and contribute to plaque stability. In addition, a recentstudy demonstrates that hypercoagulability in transgenic micepromotes plaque stability.27
At the same time, the activity of coagulation proteasescontributes to local inflammation and angiogenesis, andtherefore the latter will eventually prevail over processes suchas proliferation, thus compromising plaque stability. Thisproinflammatory state of the evolving plaque, including
828 Circulation August 24, 2010
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
increased apoptosis of SMC, gradual protein loss, and en-hanced angiogenesis, will herald plaque evolution and greatervulnerability. Hence, EAL may be more stable because ofmore clotting activity, whereas SAAL may be more vulner-able because of instability. In the case of a plaque rupture,even relatively small amounts of TF and other proteins maystill be highly thrombogenic, precipitating thrombus forma-tion and cardiovascular events.
In conclusion, our findings provide substantial new dataillustrating the close involvement of coagulation proteins inthe entire process of atherogenesis. Whereas in the earlylesions essentially all coagulation proteins, including thosefrom the contact/intrinsic system, are readily detectable(possibly supporting plaque stability), on transformation toadvanced lesions the amount and activity of these proteinsdiminish. The loss in coagulation activity, possibly due toincreased inflammatory pressure, may reduce plaque stabilityand contribute to the risk of plaque rupture. These resultspoint to various and specific functions of coagulation proteinsin regulating progression of atherosclerosis and may providenovel insights into the genesis of atherothrombosis. Thesedata also suggest ways to modulate atherogenesis and possi-bly reduce atherosclerosis that may eventually be clinicallyuseful. The fact that new specific anticoagulant agents arebeing clinically tested underscores the necessity of furtherstudies in this area.
AcknowledgmentsDrs Heeneman and Daemen participate in the European VascularGenomics Network (http://www.evgn.org), a Network of Excellencesupported by the European Commission’s Sixth Framework Programfor Research Priority 1 (Life Sciences, Genomics, and Biotechnologyfor Health; contract LSHM-CT-2003-503254). We gratefully ac-knowledge Diane Fens, Patricia Pluijmen, and Mathijs Groenewegfor their skillful help in processing and measuring the specimens.
Sources of FundingMarie Curie fellowships from the European Commission weregranted to Dr Borissoff (MEST-CT-2005-020706) and Peter Kassak(QLK5-CT-2000-6007). Evren Kilinç and Dr Winckers are spon-sored by the Netherlands Heart Foundation (grants 2006-B064 and2007-B138).
DisclosuresResearch grants as principal investigators from the Center forTranslational Molecular Medicine, the Netherlands, were granted toDrs Hackeng, Daemen, and ten Cate; from the Netherlands HeartFoundation were granted to Drs Hackeng and ten Cate; and fromFonds Economische Structuurversterking were granted to Dr Hack-eng. Dr ten Cate received honoraria as a consultant to BoehringerIngelheim GmbH. Dr Hackeng received honoraria from ACS-Biomarker B.V. The other authors report no conflicts.
References1. Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med. 1999;
340:115–126.2. Levi M, van der Poll T, Buller HR. Bidirectional relation between inflam-
mation and coagulation. Circulation. 2004;109:2698–2704.3. Esmon CT. The interactions between inflammation and coagulation. Br J
Haematol. 2005;131:417–430.
4. Wilcox JN, Smith KM, Schwartz SM, Gordon D. Localization of tissuefactor in the normal vessel wall and in the atherosclerotic plaque. ProcNatl Acad Sci U S A. 1989;86:2839–2843.
5. Wilcox JN, Noguchi S, Casanova J. Extrahepatic synthesis of factor VIIin human atherosclerotic vessels. Arterioscler Thromb Vasc Biol. 2003;23:136–141.
6. Pyo RT, Sato Y, Mackman N, Taubman MB. Mice deficient in tissuefactor demonstrate attenuated intimal hyperplasia in response to vascularinjury and decreased smooth muscle cell migration. Thromb Haemost.2004;92:451–458.
7. Versteeg HH, Ruf W. Emerging insights in tissue factor-dependent sig-naling events. Semin Thromb Hemost. 2006;32:24–32.
8. Bini A, Fenoglio JJ Jr, Mesa-Tejada R, Kudryk B, Kaplan KL. Identifi-cation and distribution of fibrinogen, fibrin, and fibrin(ogen) degradationproducts in atherosclerosis: use of monoclonal antibodies. ArteriosclerThromb Vasc Biol. 1989;9:109–121.
9. Smith EB, Thompson WD. Fibrin as a factor in atherogenesis. ThrombRes. 1994;73:1–19.
10. Wilcox JN, Nelken NA, Coughlin SR, Gordon D, Schall TJ. Localexpression of inflammatory cytokines in human atherosclerotic plaques.J Atheroscler Thromb. 1994;1(suppl 1):S10–S13.
11. Laszik ZG, Zhou XJ, Ferrell GL, Silva FG, Esmon CT. Down-regulationof endothelial expression of endothelial cell protein C receptor andthrombomodulin in coronary atherosclerosis. Am J Pathol. 2001;159:797–802.
12. Borissoff JI, Spronk HM, Heeneman S, ten Cate H. Is thrombin a keyplayer in the “coagulation-atherogenesis” maze? Cardiovasc Res. 2009;82:392–403.
13. Bea F, Kreuzer J, Preusch M, Schaab S, Isermann B, Rosenfeld ME,Katus H, Blessing E. Melagatran reduces advanced atherosclerotic lesionsize and may promote plaque stability in apolipoprotein E–deficientmice. Arterioscler Thromb Vasc Biol. 2006;26:2787–2792.
15. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons fromsudden coronary death: a comprehensive morphological classificationscheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol. 2000;20:1262–1275.
16. Ragni M, Golino P, Cirillo P, Scognamiglio A, Piro O, Esposito N,Battaglia C, Botticella F, Ponticelli P, Ramunno L, Chiariello M. Endog-enous tissue factor pathway inhibitor modulates thrombus formation in anin vivo model of rabbit carotid artery stenosis and endothelial injury.Circulation. 2000;102:113–117.
17. St Pierre J, Yang LY, Tamirisa K, Scherrer D, De Ciechi P, Eisenberg P,Tolunay E, Abendschein D. Tissue factor pathway inhibitor attenuatesprocoagulant activity and upregulation of tissue factor at the site ofballoon-induced arterial injury in pigs. Arterioscler Thromb Vasc Biol.1999;19:2263–2268.
18. Badimon JJ, Lettino M, Toschi V, Fuster V, Berrozpe M, Chesebro JH,Badimon L. Local inhibition of tissue factor reduces the thrombogenicityof disrupted human atherosclerotic plaques: effects of tissue factorpathway inhibitor on plaque thrombogenicity under flow conditions.Circulation. 1999;99:1780–1787.
19. Martin K, Weiss S, Metharom P, Schmeckpeper J, Hynes B, O’SullivanJ, Caplice N. Thrombin stimulates smooth muscle cell differentiationfrom peripheral blood mononuclear cells via protease-activatedreceptor-1, RhoA, and myocardin. Circ Res. 2009;105:214–218.
20. Mackman N. Role of tissue factor in hemostasis, thrombosis, and vasculardevelopment. Arterioscler Thromb Vasc Biol. 2004;24:1015–1022.
21. Barbasz A, Kozik A. The assembly and activation of kinin-formingsystems on the surface of human U-937 macrophage-like cells. BiolChem. 2009;390:269–275.
22. Maas C, Govers-Riemslag JW, Bouma B, Schiks B, Hazenberg BP,Lokhorst HM, Hammarstrom P, ten Cate H, de Groot PG, Bouma BN,Gebbink MF. Misfolded proteins activate factor XII in humans, leading tokallikrein formation without initiating coagulation. J Clin Invest. 2008;118:3208–3218.
23. Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s.Nature. 1993;362:801–809.
24. Kockx MM, De Meyer GR, Muhring J, Bult H, Bultinck J, Herman AG.Distribution of cell replication and apoptosis in atherosclerotic plaques ofcholesterol-fed rabbits. Atherosclerosis. 1996;120:115–124.
25. Shah PK, Falk E, Badimon JJ, Fernandez-Ortiz A, Mailhac A,Villareal-Levy G, Fallon JT, Regnstrom J, Fuster V. Human monocyte-
Borissoff et al Coagulation Markers in Atherosclerosis 829
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
derived macrophages induce collagen breakdown in fibrous caps of ath-erosclerotic plaques: potential role of matrix-degrading metallopro-teinases and implications for plaque rupture. Circulation. 1995;92:1565–1569.
26. Miller DT, Ridker PM, Libby P, Kwiatkowski DJ. Atherosclerosis: thepath from genomics to therapeutics. J Am Coll Cardiol. 2007;49:1589–1599.
27. Seehaus S, Shahzad K, Kashif M, Vinnikov IA, Schiller M, Wang H,Madhusudhan T, Eckstein V, Bierhaus A, Bea F, Blessing E, Weiler H,Frommhold D, Nawroth PP, Isermann B. Hypercoagulability inhibitsmonocyte transendothelial migration through protease-activatedreceptor-1-, phospholipase-Cbeta-, phosphoinositide 3-kinase-, and nitricoxide-dependent signaling in monocytes and promotes plaque stability.Circulation. 2009;120:774–784.
CLINICAL PERSPECTIVEApart from their well-established role in coagulation, several hemostatic factors (eg, tissue factor/activated factor VIIcomplex, activated factor X, thrombin) have been reported to evoke multiple proatherogenic events on a wide range ofarterial wall constituents. While exploring the presence and distribution of all coagulation proteins in both early andadvanced human atherosclerotic plaques, we found a colocalization of key procoagulant proteins with smooth muscle cellsand macrophages, thus suggesting an active cell-based coagulation network within the atherosclerotic plaque. Furthermore,we provide new evidence pointing toward local synthesis of several coagulation factors within the atherosclerotic vesselwall. The principal finding of this study, indicating enhanced procoagulant activity of early atherosclerotic plaques versusstable advanced plaques, suggests a role for the hemostatic proteins and hypercoagulability in regulating the onset andprogression of atherosclerosis. These findings may become clinically relevant in the new era of selective oralanticoagulants, in which such agents may have effects on the complex process of atherosclerosis beyond their directantithrombotic action.
830 Circulation August 24, 2010
by guest on October 14, 2011http://circ.ahajournals.org/Downloaded from
