Recanalization of arterial thrombus, and inhibition with β-radiation in a new murine carotid...

BASIC RESEARCH STUDIES

Recanalization of arterial thrombus, and inhibitionwith �-radiation in a new murine carotid occlusionmodel: mRNA expression of angiopoietins,metalloproteinases, and their inhibitorsJean Raymond, MD,a Vickie Lebel, MSc,a Christelle Ogoudikpe, MSc,a Annick Metcalfe, MSc,a MiguelChagnon, MSc,b and Olivier Robledo, PhD,a Montreal, Quebec, Canada

Background: Recanalization is an important physiologic phenomenon because it can efficiently reestablish circulation afterthrombosis. We attempted to characterize molecular events related to recanalization or organization of arterial thrombusin a new murine model by studying genes reported to be involved in angiogenesis or neointima formation.Methods: Platinum coils, radioactive phosphorus 32 coils or not, were implanted in the carotid artery in mice to causethrombotic occlusion. The outcome of the occlusion was followed up with transmyocardial angiography and pathologicanalysis at 2, 6, or 15 days. Angiographic results were compared with the Pearson �2 test. Messenger RNA expression ofvon Willebrand factor (vWF); smooth muscle �-actin (SMA�); platelet endothelial cell adhesion molecule-1 (PECAM-1);vascular endothelium cadherin (VE-Cad); endothelial nitric oxide synthase (eNOS); vascular cell adhesion molecule-1(VCAM-1); tumor necrosis factor alpha (TNF-�); matrix metalloproteinase (MMP-9, MMP-12, and MMP-14), andtissue inhibitors of MMPs (TIMPs: TIMP-1, TIMP-2, TIMP-3, TIMP-4); angiopoietins (Ang-1, Ang-2); and receptorsTie-1 and Tie-2, were analyzed with reverse transcriptase polymerase chain reaction 2, 6, and 15 days after surgery. Levelsof mRNA expression were compared with analysis of variance and the Student t test.Results: Carotid arteries implanted with nonradioactive 0.015-caliber coils were occluded in 84% of arteries on day 2, but inonly 57% of arteries on day 15, which confirms that recanalization occurred in this model. Arteries implanted with0.015-caliber 32P coils did not become recanalized, and 100% were occluded on day 15 (n � 13; P � .006). Recanalization wasassociated with endothelial-like cell-lined channels, whereas persistent occlusion was caused by complete filling of the lumenwith conjunctive tissue. Coil occlusion, with or without recanalization, was followed by decreased expression of vWf, VE-Cad,eNOS, VCAM-1, MMP-2, TIMP-1, and TIMP-2; stable expression of PECAM-1, SMA�, and TIMP-3; and overexpressionof Ang-1 and Ang-2, MMP-9, MMP-14, and TIMP-4. Statistically significant differences when arteries were implanted with32P coils included decreased expression of TIMP-4 (P � .011) and increased expression of MMP-9 (P � .02).Conclusion: Recanalization and organization of arterial thrombus is associated with expression of genes involved inangiogenesis and neointima formation. Recanalization can be prevented with �-radiation, but molecular mechanismsremain to be refined. (J Vasc Surg 2004;40:1190-8.)

Clinical Relevance: A better understanding of molecular mechanisms involved in angiogenesis has permitted itsregulation as a new option in treatment of various diseases. Inhibition of angiogenesis may help control diseases such ascancer, arthritis, or diabetes retinopathy. On the other hand, stimulation of angiogenesis may palliate conditionsassociated with insufficient blood supply, such as ischemic heart disease or critical limb ischemia. Yet little is knownregarding recanalization (to be differentiated from thrombolysis), a cellular process that occurs concurrently withthrombus “organization.” Recanalization is an important physiologic phenomenon because it can efficiently reestablishantegrade circulation after thrombosis both in veins and in arteries, and could be modulated for therapeutic purposes.Thus our efforts at better understanding of mechanisms involved in recanalization could be used, in addition to itspromotion to recover flow after thrombotic occlusions, to prevent its occurrence after endovascular interventions
designed to permanently occlude aneurysms.
Murine vascular models have permitted considerableprogress in understanding of mechanisms involved in arte-rial thrombosis, neointima formation, and development of

From the CHUM Research Centre, Notre-Dame Hospital,a and the De-partment of Mathematics and Statistics, University of Montreal.b

This work was made possible through grants for the Canadian Institutes ofHealth Research (CIHR) Diabete Quebec. O.R. is a scholar from theFonds de la Recherche en Sante du Quebec (FRSQ).

Competition of interest: none.Reprint request: Olivier Robledo, PhD, Interventional Neuroradiology

Research Laboratory, CHUM Research Center, Notre-Dame Hospital,1560 Sherbrooke E, Ste M-8203, Montreal, Que, H2L 4M1 Canada.(e-mail: [email protected]).

0741-5214/$30.00Copyright © 2004 by The Society for Vascular Surgery.
doi:10.1016/j.jvs.2004.09.023
1190

collateral circulation by angiogenetic and more recentlyvasculogenetic pathways.1-4 Yet little is known regardingrecanalization, a cellular process that occurs with thrombus“organization”5-7 and described by pathologists for de-cades.8 Recanalization is an important physiologic phe-nomenon because it can efficiently reestablish antegradecirculation after thrombosis both in veins and in arteries,and thus it could be modulated for therapeutic purposes.

Recanalization is routinely found after platinum coilocclusion of canine arteries, and could be involved inrecurrences after endovascular treatment of human intra-cranial aneurysms.9-12 Deficient recanalization has notbeen described, except perhaps in transgenic mice,13 but
it is conceivable that it could be impaired in diseases

JOURNAL OF VASCULAR SURGERYVolume 40, Number 6 Raymond et al 1191

associated with abnormal angiogenetic mechanisms,such as diabetes.14

A better understanding of molecular events related torecanalization could lead to interventions designed to pro-mote or prevent its occurrence. If inhibition of recanaliza-tion could improve results of neurovascular interventions,its promotion may find a role in the subacute managementof occlusive vascular disorders such as lower limb ischemia.

We have previously described canine models that werehelpful in assessing neurovascular devices currently used inclinical practice, but they are limited in terms of biomolec-ular tools essential to exploration of mechanisms of recan-alization.9-12 Development of a new murine arterial coilocclusion model is presented. This model was designed toenable future in-depth studies of recanalization mecha-nisms and to assess the competence of this phenomenon inmodels of atherosclerosis or diabetes, in transgenic animalsor cell grafting experiments.1,7,13

Recanalization, but not thrombus organization, canreliably be prevented in canine arteries by in situ �-radiationat low doses.9-12 By using radioactive or nonradioactivecoils in this murine model, we confirmed inhibition ofrecanalization with �-radiation and attempted to identifymolecular events related to thrombus organization, with orwithout recanalization. Many mechanistic avenues couldhave been explored, but because organization and recana-lization of thrombus may share conceptual mechanismswith neointima formation and angiogenesis, respectively,we chose to address first messenger RNA expression ofsome endothelial and neointimal cell markers, and growthfactors and metalloproteinases reportedly involved in thesephenomena (Table).

MATERIAL AND METHODS

Experimental animals and surgical procedure. Twohundred thirteen C57BL/6 wild-type mice (Charles River)weighing 18 to 20 g were administered preoperative andpostoperative analgesia with 0.03 mg/kg of buprenorphinehydrochloride before and for 2 days after surgery. Allexperimental procedures were approved by the institutionalanimal care committee in accordance with guidelines of theCanadian Council on Animal Care. Animals were anesthe-tized with isoflurane. A midline cervical incision was madeto access the right carotid artery, which was carefully dis-sected. A drop of bupivacaine (0.25%) was locally applied.The right carotid artery was exposed and elevated with 26-0 polypropylene sutures. A 25-gauge needle bent at 90degrees was used to perform the arteriotomy, and astretched platinum coil of 0.010 or 0.015 caliber, 2.7 � 0.6mm long (Target Therapeutics, Boston Scientific), radio-active or not radioactive, was introduced into the rightcarotid artery (Fig 1). Because 0.015-caliber coils led tomore reliable initial occlusions (see Results), all furtherexperiments and reverse transcriptase–polymerase chain re-action (RT-PCR) analyses were performed with these coils.The sutures were removed and the arteriotomy site con-trolled with local compression. The incision was closed
with 6-0 polypropylene sutures.
Ion implantation of coils. Phosphorus 32 coils wereproduced as previously described10 on a dedicated radioac-tive ion implanter at Université de Montréal. 32P waspurchased from Perkin-Elmer Life Sciences. Coils, a kindgift from Target Therapeutics, were measured after implan-tation, identified, sterilized, and inserted into Plexiglasboxes. Coils were sectioned every 2 to 3 mm, and for eachsegment to be implanted into the carotid artery the 32Pactivity in adjacent segments was counted, as described.10

Mean coil linear activity was 0.023 � 0.018 �Ci/mm, andmean volumetric activity inserted into carotid arteries was0.23 � 0.16 �Ci/mm3.

Transmyocardial angiography and sacrifice. Angi-ography was performed at 2, 6, or 15 days after surgery.Mice were injected with a lethal dose of pentobarbital(Somnotol; MTC Pharmaceuticals, Cambridge, Ontario,Canada) (65 mg/mL; 120 mg/kg intraperitoneally) justbefore angiography. A 1-mL syringe filled with radiopaquecontrast medium (iodixanol [Visipaque], 320 mg/mL)and a 22-gauge needle were used. The chest was openedand the needle inserted in the left ventricle for injection. Aleft lateral oblique projection was used in all animals, anddigital subtraction images were interpreted with blindedassessment (Fig 1). Occlusion was defined as the absence ofanterograde blood flow through the coiled arterial seg-ment; any antegrade opacification was sufficient to label theartery as recanalized.

Pathologic technique. In brief, right and left carotidarteries in each mouse were carefully dissected, fixed in 10%neutral buffered formalin, and embedded in paraffin. Incertain specimens the coils were carefully removed under anoperating microscope after 24 hours of fixation. Alterna-tively, the carotid arteries were left in situ, and the head andneck were fixed en bloc before embedding and axial sec-tioning. Specimens, including the coil, were cut with anExakt diamond saw after dehydratation and embedding inTechnovit 7200 resin (Heraeaus Kulzer). Slides werestained with hematoxylin-eosin, Verhoeff–van Gieson, orhematoxylin-phloxine-saffron, as described.9

RNA isolation and RT-PCR analysis. Isolation oftotal cellular RNA was performed with the Micro-to-Midisystem (Invitrogen) according to the manufacturer’s in-structions. Aliquots of total cellular RNA (500 ng) wereused for first-strand complementary DNA synthesis in 20�L of reaction volume with Superscript II reverse transcrip-tase (Invitrogen). After determining the linear range ofRT-PCR for each of the target genes, amplification of thegene under investigation was carried out with primers(Table) and chosen in 2 different exons to distinguishgenomic contamination. PCR amplifications were per-formed with platinium Taq DNA polymerase (Invitrogen)according to the manufacturer’s instructions, on an Epen-dorf Mastercycler gradient, with the following program:step 1, 94°C for 1 minute; step 2, 52°C to 64°C for 1minute; and step 3, 72°C for 1 minute (Table). Forty cycleswere performed for amplification of genes of interest, and30 cycles for �-actin. The amplification for each gene was in
the linear curve. PCR products were visualized on 1.5% aga-

thase;ode; T

JOURNAL OF VASCULAR SURGERYDecember 20041192 Raymond et al

rose gel stained with ethidium bromide and ultraviolet tran-sillumination. Quantitative analysis was carried out with acomputerized densitometric imager (ImageQuant) to obtaingenes to �-actin ratios. For each gene, results from 7 to 24specimens per time point were used for analysis.

Statistical analysis. Angiographic occlusion rates atdifferent times with radioactive or nonradioactive coils ofvarious caliber were compared with the Pearson �2 test.mRNA expression levels, analyzed with densitometry, werecompared with the Student t test and analysis of variance,followed, if necessary, with the Tukey test. To control thefalse discovery rate, the Benjamini and Hochberg correc-tion was used. For all statistical tests, P � .05 was consid-ered statistically significant.

RESULTS

Surgical, angiographic, and pathologic findings

Surgical mortality rate was minimal, 8% overall. Carotidocclusion, with or without recanalization, did not lead toneurologic deficits in any animal. The arteriotomy itself didnot lead to carotid occlusion, and was followed by minimalparietal changes, a small local hemorrhagic exudate withinflammatory cells, at histopathologic analysis (n � 3; datanot shown).

The surgical technique was modified to obtain a reli-able occlusion at 2 days. The use of 0.010-caliber filamentswas followed by occlusion on day 2 in only 53% of arteries,compared with 84% with 0.015-caliber filaments (P �

Reverse transcriptase–polymerase chain reaction primer seq

Gene Tm NT (bp) Forward

Endothelialmarkers

vWF 64 1097 5=-GCGCATCCGCGTGeNOS 64 834 5=-AGCACCCTGGCAGVCAM-1 60 400 5=-ATGCCTGTGAAGAVECAD 64 520 5=-TTGCCCAGCCCTAPECAM 52 452 5=-AGGGGACCAGCTG

Smoothmusclecellmarkers

SMA� 52 240 5=-GAGAAGCCCAGCC

MMPs MMP-2 58 705 5=-CTTGCAGGAGACAMMP-9 58 380 5=-CCATGAGTCCCTGMMP-14 58 643 5=-ACATTCTAACGATC

Inhibitorsof MMPs

TIMP-1 58 655 5=-CTTTGCATCTCTGTIMP-2 58 584 5=-TGCAGCTGCTCCCTIMP-3 58 392 5=-AACTCCGACATCGTIMP-4 58 491 5=-TCGGCTCTAGTGA

Growthfactors

VEGF 60 620-548-488

5=-GGATCCATGAACT

Ang-1 64 580 5=-CAGTGGCTGCAAAAng-2 64 666 5=-CACACTGACCTTCTNF-a 58 450 5=-CAGCCTCTTCTCA

Receptors Tie-1 64 300 5=-ACCCACTACCAGCTie-2 52 414 5=-CCTTCCTACCTGC

Reporter �-actin 52/64

595 5=-CATGGATGACGAT

Ang, Angiopoietin; VEGF, vascular endothelial growth factor; PECAM, plaVCAM, vascular cell adhesion molecule; eNOS, endothelial nitric oxide synnecrosis factor; MMP, matrix metalloproteinase; Tie, transient ischemic epis

.035; n � 15 each). Thus 0.015-caliber filaments were used

for all further experiments. Carotid arteries implanted withnonradioactive 0.015-caliber coils were occluded in 57%(n � 30) of arteries on day 15, which confirmed thatrecanalization occurred in this model. When the coil wasradioactive recanalization was prevented, as shown by oc-clusion in 100% of 13 arteries (P � .006).

Pathologic findings were degraded by a number ofartifacts caused by the presence of the metallic foreign bodyor by the dissection performed to remove the coil beforesectioning. Planimetric measurements were not performed,but complete occlusions were associated with obvious de-crease in the caliber of the carotid artery beyond the seg-ment bearing the coil. Arteries that remained occluded 15days after implantation of radioactive or nonradioactivecoils showed a carotid lumen filled with neointimal tissuecomposed of collagen and myofibroblasts; arteries im-planted with nonradioactive coils were recanalized byspaces lined with endothelial-like cells and surrounded by asimilar neointimal tissue (Fig 2).

mRNA expression after occlusion, thrombusorganization, or recanalization

Evolution of mRNA expression with time after coilocclusion. Coil occlusion was followed by significant dif-ferences, as compared with control carotid arteries, in levelsof expression of endothelial markers (VE-Cad, eNOS);matrix metalloproteinase (MMP-9, MMP-14) and tissue

es for selected genes

r Reverse primer

GTGG-3= 5=-GGCAGTTGCAGACCCTCCTTG-3=AGTCC-3= 5=-GGTAGCGTTGCTGATCC-3=TCGC-3= 5=-GTGACCGGCTTCCCAACCTC-3=ACCTAAAG-3= 5=-ACCACCGCCCTCCTCATCGTAAGT-3=ATTAGG-3= 5=-AGGCCGCTTCTCTTGACCACTT-3=CG-3= 5=-CTCTTGCTCTGGGCTTCA-3=

TCTGG-3= 5=-TTAAGGTGGTGCAGGTATCTGG-3=G-3= 5=-AGTATGGATGTTATGATG-3=CCG-3= 5=-TCACTGCCCATGAATGACCC-3=CTGG-3= 5=-AAGTAGACAGTGTTCAGGC-3=

TGCAC-3= 5=-TTATGGGTCCTCGATGTCAAG-3=CCGGGCC-3= 5=-GGAGAGCATGTCGGTCCAGAGACA-3=GGCC-3= 5=-CCTCGGTACCAGCTGCAGATG-3=GCT-3= 5=-GAATTCACCGCCTCGGCTTGTC-3=

TGA-3= 5=-TCTGCACAGTCTCGAAATGG-3=ACT-3= 5=-TGGTGTCTCTCAGTGCCTTG-3=TGC-3= 5=-TTGACTTTCTCCTGGTATGAGATA-3=

ATGT-3= 5=-ATCGTGTGCTAGCATTGAGG-3== 5=-CCACTACACCTTTCTTTACA-3=

CTGCGC-3= 5=-GCTGTCGCCACGCTCGGTCAGGAT-3=

dothelial cell adhesion molecule; VECAD, vascular endothelium cadherin;vWF, von Willebrand factor; SMA�, smooth muscle �-actin; TNF, tumorIMP, tissue inhibitor of matrix metalloproteinases.

uenc

prime

GCAAAGTGGGCACACAGT

AGTGCACCT

GCATCGGTGATTACGTTCT

AACTCCCATTCCTGGTA-3ATCG

telet en

inhibitors of MMPs (TIMP-2, TIMP-4); and growth fac-


tors (Ang-1, Ang-2, TNF-�). In addition, there were sig-nificant differences in the evolution of expression of certaingenes over time (vWf, eNOS, VCAM-1, MMP-2, TIMP-1,and Ang-2).

The expression levels of some endothelial markers, suchas VE-Cad and eNOS, were decreased at 2 and 6 days.Others were decreased, but to levels that were not signifi-cantly different from control levels (vWf, VCAM-1, Tie-1).Some genes (vWF, VCAM-1) reached control levels by 15days (Fig 3), wheres others remained low (VE-Cad,

Fig 1. Coil occlusion model and angiographic results. a, Intraop-erative photograph shows endoluminal implantation of platinumcoil (arrowheads). b, Transmyocardial angiogram of normal ca-rotid artery is shown for comparison. Transmyocardial angiograms2 days (c, e) and 15 days (d, f) after surgical implantation of notradioactive (c, d) or radioactive (e, f) platinum coils show occlu-sion on day 2 with both types of coils (c, e). Artery remainsoccluded at 15 days when coil is radioactive (arrows in f), butrecanalizes when coil is not radioactive (d).

eNOS). Smooth muscle cell �-actin (SMA�) expression

remained unchanged throughout experiments, whereasTNF-� was significantly increased at 2 days (data notshown).

Expression of MMP-9 and MMP-14 was elevated ascompared with control arteries, whereas MMP-2 expres-sion was low early on, at 2 days, but increased to reachcontrol levels by 15 days (P � .015; Fig 4). Expression ofTIMP-2 was decreased at all times, most significantly at 6days (P � .001). TIMP-1 expression increased over time(P � .011 between 2 and 15 days), but remained withincontrol levels. There was no significant modulation ofTIMP-3. There was significant overexpression of TIMP-4,maximal at 2 days (P � .011; Fig 4).

Ang-1 showed 8-fold to 12-fold increased expressioncompared with controls (P � .000), whereas Ang-2 expres-sion was elevated (P � .047) only on day 6 (Fig 5).

mRNA expression after radioactive coil occlu-sion. There was no significant difference in evolution overtime in mRNA expression of vWF, PECAM-1; VE-Cad;eNOS; SMA�; TNF-�; MMP-2 or MMP-14; TIMP-1,TIMP-2, or TIMP-3; Tie-1 or Tie-2; and Ang-1 or Ang-2between arteries implanted with 32P or nonradioactivecoils. Statistically significant differences between arteriesimplanted with 32P or nonradioactive coils included de-creased expression of TIMP-4 (P � .011) and increased

Fig 2. Coil occlusion of carotid artery: pathologic findings. Axialsections of normal carotid artery (a) and of carotid artery 2 days (b)or 15 days (c, d) after coil occlusion (arrow in b). Artery isrecanalized (asterisks in c) when coil was not radioactive, whereaslumen was completely filled with conjunctive tissue (d) when coilwas radioactive. c and d, Images are of sections immediately abovecoil. (a, b, Hematoxylin-eosin stain; c, d, Movat pentachromestain; original magnification 100 [a] or 50 [b-d]).

expression of MMP-9 (P � .02).


DISCUSSION

Recanalization is constant after coil occlusion of caninearteries and can be inhibited with �-radiation.9,10 Arterialrecanalization after coil occlusion is a well-known clinicalphenomenon. It has also been documented in many spe-cies, including dogs, pigs, rabbits, and now mice.9,10 It is acellular process that occurs in delayed fashion,6-10 withindays, and thus should be distinguished from thrombolysis.

Recanalization in a murine carotid artery coil-occlusionmodel can also be inhibited by in situ �-radiation. Themurine model was more difficult to standardize, and angio-graphic results were less reliable than in the canine model,in which occlusion was found in 100% of arteries withinhours of coil occlusion; thrombotic occlusion persisted for6 to 10 days, but was constantly followed by recanalizationby 15 days when coils were nonradioactive.9,10,12 We con-centrated our investigation on the first 2 weeks because, ifrecanalization was prevented during this early period itwould not occur later, when animals were followed up for 3months.10

The small size of the murine carotid artery, comparedwith the caliber of the metallic coil, could have been an

Fig 3. Gene expression of endothelial markers. Reverse transcrip-tase–polymerase chain reaction analysis of carotid arteries im-planted with radioactive coils (white bars) or nonradioactive coils(black bars). Results at 2, 6, or 15 days are expressed as ratiorelative to expression found in normal carotid arteries (controlarteries [CTL], gray bars). Note decreased expression of mostgenes at 2 and 6 days with both types of coils. vWF, von Willebrandfactor; PECAM, platelet endothelial cell adhesion molecule;VCAM-1, vascular cell adhesion molecule-1; VE-CAD, vascularendothelium cadherin; eNOS, endothelial nitric oxide synthase.

important factor to explain the decreased efficacy of the

recanalization phenomena in mice. Less reliable initial oc-clusion and delayed recanalization rates may introducevariability in biomolecular analyses, as well as difficulties ininterpretation of results when animals are sacrificed beforerecanalization can be documented at angiography orpathologic study, when it is presumed that causal mecha-nisms are active (�6-10 days).9-12 Another weakness of themodel is the presence of the metallic foreign body, whichlimits pathologic and immunohistologic studies. Manytechnical details remain to be adjusted to maximize patho-logic techniques. We have previously shown in larger spe-cies that recanalization consisted of formation of endothe-lial cell–lined longitudinal channels crossing the throm-botic occlusion.9,10 Although we could not prove the en-dothelial nature of the cells bordering the channels, weassume recanalization in this murine model is of a similarnature.

There has been little interest in recanalization untilrecently. A vena cava thrombus organization model hasbeen described in rats and mice. This model enabled assess-ment of the effects of local administration of vascular en-dothelial growth factor, and the potential role of circulatingcells, monocytes, or progenitor cells in venous recanaliza-tion after thrombotic occlusion.6,7,15-17 Increased organi-zation with administration of monocyte chemotactic pro-tein-1 and increased neovascularization with interleukin-8have also been found in the same model.6,17 Biologicprocesses involved in venous and arterial recanalization maydiffer, however. In the absence of urokinase plasminogenactivator (u-PA), cellular invasion of venous thrombi, andconsequently organization and recanalization are defi-cient.13

Cellular and molecular mechanisms responsible forthrombus organization and recanalization remain poorlydefined. Diverse hypotheses have been reviewed by Moldo-van and Asahara.5 Organization and recanalization involveinvasion of the clot by parietal or circulating mononuclearor progenitor cells forming channels, endothelialization,myofibroblastic invasion, deposition of collagen, and con-traction of the provisional matrix. Thus there may becellular and molecular analogies between organization andneointima formation and between angiogenesis and recan-alization. One hypothesis proposes a key role for mono-cytes, which may drill metalloelastase-positive tunnelswithin the fibrin clot, an event potentially followed byseeding of circulating progenitor cells, at least in angiogen-esis, but also possible in recanalization.5,15,18

The role of progenitor cells has been emphasized in anexperiment in which deficient organization and recanaliza-tion in uPA-deficient mice was rescued by infusion ofwild-type “progenitor” cells.7 Bone marrow cells make avarying contribution to neointima formation, dependingon animal models,3,19,20 and various circulating cellularlineages may be involved in organization or recanaliza-tion.6,16,18-20 The controversy regarding the role, or eventhe identity, of circulating progenitor cells is underlined byreports that propose they may be derived from monocytes
or bone marrow monocyte lineage cells,21,22 and may be

betw


more active by secreting growth factors than by a directcellular participation in construction of the “new ves-sels.”23

Endothelial denudation before coil implantation couldprevent recanalization.11,12 Thus in the current murinemodel we studied mRNA expression of endothelial markers(vWf, Tie-2), adhesion molecules (VE-Cad, VCAM-1,PECAM-1), growth factors (Ang-1, Ang-2), and MMPspreviously shown to be essential to endothelial invasionand in vitro vessel formation in fibrin matrices, such asMMP-2, MMP-9, and MMP-14.24-27 The initial decreasedexpression of endothelial markers and adhesion molecules,followed by recovery, may be interpreted as a sign ofendothelial damage caused by coil occlusion followed byregeneration. The stable expression of Tie-2, with no initialdecrease in expression, may be surprising in this context.Progenitor circulating or bone marrow cells may, however,express “endothelial markers” both in vitro and in vivo, anduse similar MMPs to invade matrices, at least in vitro.28

Because monocytes and progenitor cells, including �-actinmural precursor cells, may reach the thrombus early on andcan express markers traditionally attributed to endothelialcells, interpretation of these experiments at the cellular levelshould be done with caution.29

MMPs have a central role in degrading extracellular andbasement membranes, a necessary step in angiogenesis,which could also be important to thrombus organizationand recanalization. The hypothesis of a milieu that favorscellular migration into the thrombus is supported by thesubstantial increase in MMP-9 and MMP-14, but stable ordecreased expression of their inhibitors (TIMP-1, TIMP-2,TIMP-3). To our surprise, there was no significant differ-ence between arteries occluded with nonradioactive coils,

Fig 4. Gene expression of matrix metalloproteinasestranscriptase–polymerase chain rection analysis of MMradioactive coils (white bars) or nonradioactive coils (blacto expression found in normal carotid arteries (controMMP-9 (P � .003) and TIMP-4 (P � .001) expression

which tend to recanalize, and those treated with 32P coils,

which remained occluded at all times for most MMPsstudied. The differential expression of MMP-9 is unlikely toexplain inhibition of recanalization with radiation, becausethe observed difference is difficult to reconcile with knownproperties of MMPs and was most significant at 15 days,perhaps too late to be causal. MMPs have other propertiesthat we could link to organization and recanalization.MMP-2 and MMP-9 have a role in smooth muscle cellmigration and contraction of collagen matrices.30 Thisproperty of smooth muscle cells or myofibroblasts, pre-sumed essential to thrombus organization, may also beinvolved in recanalization.

A contrasting finding was the decreased level ofMMP-2 with both types of coils. Downregulation ofMMP-2 expression has been described in hypoxic condi-tions, which could have been present after coil occlusion, inboth endothelial and bone marrow mesenchymal cells,which may have angiogenic potential.31,32 Decreased ex-pression of eNOS, usually elevated with hypoxia, does notsupport this hypothesis.33

MMP activities are counteracted by TIMPs, which canbe used to control neointima formation.34,35 TIMP-1 andTIMP-3 were little affected whether the coil was radioac-tive or not. Decreased expression of TIMP-2 could con-ceivably favor cellular invasion and recanalization, and maybe related to increased Ang-1.36 TIMP-4, expressed aftervascular injury and associated with collagen accumulation,was significantly elevated only in arteries treated with non-radioactive coils (Fig 4). Because TIMP-4 reduces smoothmuscle cell migration, it is tempting to speculate thatorganization by smooth muscles cells was favored overrecanalization by endothelial or progenitor cells in radioac-tive arteries.37 Because the functions of TIMP-4 remain

Ps) and tissue inhibitors of MMPs (TIMPs). Reversed TIMP expression in carotid arteries implanted with). Results at 2, 6, or 15 days are expressed as ratio relativeries [CTL], gray bars). Note significant differences ineen radioactive and nonradioactive coils.

(MMP ank barsl arte

poorly documented, this finding deserves further study.


Expression of growth factors involved in angiogenesis,such as Ang-1 and Ang-2, and related receptors Tie-1 andTie-2, were also studied.38,39 Angiopoietins share sequencehomologies with fibrinogen, and cause “sprouting angio-genesis” in fibrin matrices in vitro.38 Thus they may have arole in endothelial invasion of the thrombus and potentiallyin recanalization in this model. Angiopoietins are alsoinvolved in recruitment of perivascular cells, and thus couldfavor organization.38 There was increased expression ofTie-2, but decreased expression of Tie-1, after coil occlu-sion, although these variations were not found significant(Fig 5). Ang-2 mRNA showed decreased levels on day 2when coils were radioactive; it was tempting to speculatethat decreased levels could correlate with decreased recan-alization, but the difference did not reach statistical signif-icance. Arteries treated with radioactive or nonradioactivecoils showed 10-fold increased expression of Ang-1, agrowth factor that should favor “angiogenic phenomena.”Ang-1 has also been shown to associate with stabilization ofvessels, which is consistent with organization of throm-bus.40 We were unable to document vascular endothelialgrowth factor mRNA expression with the primers that weused. A significant difference in expression of genes in-

Fig 5. Gene expression of angiopoietins (Ang) and recanalysis of Ang-1 and Ang-2, and Tie-1 and Tie-2 expresbars) or nonradioactive coils (black bars). Results at 2, 6in normal carotid arteries (control arteries [CTL], graoccluded with radioactive coil occurred only on day 2 an

volved in angiogenesis between arteries treated with 32P or

nonradioactive coils could have supported a conceptualanalogy between angiogenesis and recanalization.

Conversely, the increasing levels of TNF-� and stableexpression of SMA� are compatible with smooth musclecell or myofibroblastic invasion of the thrombus and sup-port a hypothetical analogy between thrombus organiza-tion and neointima formation.41,42

The absence of major differences in the evolution overtime of mRNA in most genes studied suggests that inhibi-tion of recanalization was not caused by radiation-inducedcell death.

The present work did not permit identification of keymolecular events that would offer a satisfactory, compre-hensive understanding of mechanisms responsible for re-canalization, in contrast to organization. It is also impor-tant to emphasize that RT-PCR gives no indicationregarding biologic activity. Western blot analysis was notperformed, because the scant amount of harvested materialwould have necessitated pooling of multiple (up to 5)carotid specimens per analysis. Functional data, with zy-mography, for example, could confirm the role of MMPs orTIMPs in recanalization, and in situ hybridization or im-munohistochemical methods could provide valuable infor-

s (Tie). Reverse transcriptase–polymerase chain reactionn carotid arteries implanted with radioactive coils (white5 days are expressed as ratio relative to expression founds). Decreased Ang-2 expression when carotid artery is

not reach statistical significance.

eptorsion i, or 1y bard did

mation on a possible link between protein expression and


the biologic phenomenon under study. These techniquescould be indicated once we identify a plausible mechanismand if technical difficulties involved with the study of smallspecimens that include metallic coils are overcome. Thisfirst exploration did not cover many interesting cellular ormolecular avenues, such as the potential role of platelets,monocytes, macrophages, tissue factor, serine proteases, ortheir respective inhibitors.

We plan to pursue our investigation by testing the rateof recanalization after nonradioactive coil occlusion in micelacking u-PA or MMP-9 or overexpressing MCP-1.7,30,43

We also wish to explore the competence of the recanaliza-tion phenomenon in models of diabetes or atherosclero-sis.44 The mouse strain chosen for this exploratory work canalso be used, with the proper diet, as a model of type IIdiabetes.45

CONCLUSION

Arterial recanalization can be prevented by in situ �-radiation. mRNA expression of genes involved in angio-genesis and neointima formation can also be found withrecanalization or organization of thrombus, but key cellularand molecular events remain to be better defined.

REFERENCES

1. Carmeliet P, Moons L, Collen D. Mouse models of angiogenesis,arterial stenosis, atherosclerosis and hemostasis. Cardiovasc Res 1998;39:8-33.

2. Schatteman GC, Hanlon HD, Jiao C, Dodds SG, Christy BA. Blood-derived angioblasts accelerate blood-flow restoration in diabetic mice.J Clin Invest 2000;106:571-8.

3. Tanaka K, Sata M, Hirata Y, Nagai R. Diverse contribution of bonemarrow cells to neointimal hyperplasia after mechanical vascular inju-ries. Circ Res 2003;93:783-90.

4. Doherty T, Shah P, Rajavashisth T. Cellular origins of atherosclerosis:toward ontogenetic endgame? FASEB J 2003;17:592-7.

5. Moldovan NI, Asahara T. Role of blood mononuclear cells in recanali-zation and vascularization of thrombi: past, present, and future. TrendsCardiovasc Med 2003;13:265-9.

6. Wakefield TW, Linn MJ, Henke PK, Kadell AM, Wilke CA, WrobleskiSK, et al. Neovascularization during venous thrombosis organization: apreliminary study. J Vasc Surg 1999;30:885-92.

7. Singh I, Burnand KG, Collins M, Luttun A, Collen D, Boelhouwer B,et al. Failure of thrombus to resolve in urokinase-type plasminogenactivator gene knockout mice: rescue by normal bone marrow–derivedcells. Circulation 2003;107:869-75.

8. Dible JH. Organisation and canalisation in arterial thrombosis. J PatholBacteriol 1958;75:1-7.

9. Raymond J, Leblanc P, Morel F, Salazkin I, Gevry G, Roorda S. Betaradiation and inhibition of recanalization after coil embolization ofcanine arteries and experimental aneurysms: how should radiation bedelivered? Stroke 2003;34:1262-8.

10. Raymond J, Leblanc P, Desfaits AC, Salazkin I, Morel F, Janicki C, et al.In situ beta radiation to prevent recanalization after coil embolization ofcerebral aneurysms. Stroke 2002;33:421-7.

11. Raymond J, Sauvageau E, Salazkin I, Ribourtout E, Gevry G, DesfaitsAC. Role of the endothelial lining in persistence of residual lesions andgrowth of recurrences after endovascular treatment of experimentalaneurysms. Stroke 2002;33:850-5.

12. Raymond J, Guilbert F, Metcalfe A, Gevry G, Salazkin I, Robledo O.Role of the endothelial lining in recurrences after coil embolization:prevention of recanalization by endothelial denudation. Stroke 2004;35:1471-5.

13. Schafer K, Konstantinides S, Riedel C, Thinnes T, Muller K, Dellas C,
et al. Different mechanisms of increased luminal stenosis after arterial
injury in mice deficient for urokinase- or tissue-type plasminogenactivator. Circulation 2002;106:1847-52.

14. Tepper OM, Galiano RD, Capla JM, Kalka C, Gagne PJ, JacobowitzGR, et al. Human endothelial progenitor cells from type II diabeticsexhibit impaired proliferation, adhesion, and incorporation into vascu-lar structures. Circulation 2002;106:2781-6.

15. McGuinness CL, Humphries J, Waltham M, Burnand KG, Collins M,Smith A. Recruitment of labelled monocytes by experimental venousthrombi. Thromb Haemost. 2001;85:1018-24.

16. Waltham M, Burnand KG, Collins M, McGuinness CL, Singh I, SmithA. Vascular endothelial growth factor enhances venous thrombus re-canalisation and organisation. Thromb Haemost 2003;89:169-76.

17. Humphries J, McGuinness CL, Smith A, Waltham M, Poston R,Burnand KG. Monocyte chemotactic protein-1 (MCP-1) acceleratesthe organization and resolution of venous thrombi. J Vasc Surg 1999;30:894-9.

18. Feigl W, Susani M, Ulrich W, Matejka M, Losert U, Sinzinger H.Organisation of experimental thrombosis by blood cells: evidence of thetransformation of mononuclear cells into myofibroblasts and endothe-lial cells. Virchows Arch [A] 1985;406:133-48.

19. Han CI, Campbell GR, Campbell JH. Circulating bone marrow cellscan contribute to neointimal formation. J Vasc Res 2001;38:113-9.

20. Campbell JH, Han CL, Campbell GR. Neointimal formation by circu-lating bone marrow cells. Ann N Y Acad Sci 2001;947:18-24; discus-sion 24-15.

21. Fujiyama S, Amano K, Uehira K, Yoshida M, Nishiwaki Y, Nozawa Y, etal. Bone marrow monocyte lineage cells adhere on injured endotheliumin a monocyte chemoattractant protein-1–dependent manner and ac-celerate reendothelialization as endothelial progenitor cells. Circ Res2003;93:980-9.

22. Nakul-Aquaronne D, Bayle J, Frelin C. Coexpression of endothelialmarkers and CD14 by cytokine mobilized CD34� cells under angio-genic stimulation. Cardiovasc Res 2003;57:816-23.

23. Rehman J, Li J, Orschell CM, March KL. Peripheral blood “endothelialprogenitor cells” are derived from monocyte/macrophages and secreteangiogenic growth factors. Circulation 2003;107:1164-9.

24. Burbridge MF, Coge F, Galizzi JP, Boutin JA, West DC, Tucker GC.The role of the matrix metalloproteinases during in vitro vessel forma-tion. Angiogenesis 2002;5:215-26.

25. Lijnen HR, Lupu F, Moons L, Carmeliet P, Goulding D, Collen D.Temporal and topographic matrix metalloproteinase expression aftervascular injury in mice. Thromb Haemost 1999;81:799-807.

26. Hiraoka N, Allen E, Apel IJ, Gyetko MR, Weiss SJ. Matrix metallopro-teinases regulate neovascularization by acting as pericellular fibrinol-ysins. Cell 1998;95:365-77.

27. Pepper MS. Role of the matrix metalloproteinase and plasminogenactivator-plasmin systems in angiogenesis. Arterioscler Thromb VascBiol 2001;21:1104-17.

28. Annabi B, Thibeault S, Lee YT, Bousquet-Gagnon N, Eliopoulos N,Barrette S, et al. Matrix metalloproteinase regulation of sphingosine-1-phosphate–induced angiogenic properties of bone marrow stromalcells. Exp Hematol 2003;31:640-9.

29. Schmeisser A, Graffy C, Daniel WG, RH S.Phenotypic overlap betweenmonocytes and vascular endothelial cells. Adv Exp Med Biol 2003;522:59-74.

30. Galis ZS, Johnson C, Godin D, Magid R, Shipley JM, Senior RM,Strasser RH. Targeted disruption of the matrix metalloproteinase-9gene impairs smooth muscle cell migration and �-geometrical arterialremodeling. Circ Res 2002;91:852-9.

31. Annabi B, Lee YT, Turcotte S, Naud E, Desrosiers RR, Champagne M,et al. Hypoxia promotes murine bone-marrow-derived stromal cellmigration and tube formation. Stem Cells 2003;21:337-47.

32. Ben-Yosef Y, Lahat N, Shapiro S, Bitterman H, Miller A. Regulation ofendothelial matrix metalloproteinase-2 by hypoxia/reoxygenation.Circ Res 2002;90:784-91.

33. Coulet F, Nadaud S, Agrapart M, Soubrier F. Identification of hypoxia-response element in the human endothelial nitric-oxide synthase genepromoter. J Biol Chem 2003;278:46230-40.

34. Dollery CM, Humphries SE, McClelland A, Latchman DS, McEwan
JR. Expression of tissue inhibitor of matrix metalloproteinase-1 by use


of an adenoviral vector inhibits smooth muscle cell migration andreduces neointimal hyperplasia in the rat model of vascular ballooninjury. Circulation 1999;99:3199-3205.

35. Lijnen HR, Soloway P, Collen D. Tissue inhibitor of matrix metallo-proteinases-1 impairs arterial neointima formation after vascular injuryin mice. Circ Res 1999;85:1186-91.

36. Kim I, Kim HG, Moon SO, Chae SW, So JN, Koh KN, et al. Angio-poietin-1 induces endothelial cell sprouting through the activation offocal adhesion kinase and plasmin secretion. Circ Res 2000;86:952-9.

37. Dollery CM, McEwan JR, Wang M, Sang QA, Liu YE, Shi YE. TIMP-4is regulated by vascular injury in rats. Circ Res 1999;84:498-504.

38. Teichert-Kuliszewska K, Maisonpierre PC, Jones N, Campbell AI,Master Z, Bendeck MP, et al. Biological action of angiopoietin-2 in afibrin matrix model of angiogenesis is associated with activation ofTie-2. Cardiovasc Res 2001;49:659-70.

39. Witzenbichler B, Maisonpierre PC, Jones P, Yancopoulos GD, IsnerJM. Chemotactic properties of angiopoietin-1 and -2, ligands for theendothelial-specific receptor tyrosine kinase Tie-2. J Biol Chem 1998;273:18514-21.

40. Kampfer H, Pfeilschifter J, Frank S. Expressional regulation of angio-
poietin-1 and -2 and the Tie-1 and -2 receptor tyrosine kinases during
cutaneous wound healing: a comparative study of normal and impairedrepair. Lab Invest 2001;81:361-73.

41. Rectenwald JE, Moldawer LL, Huber TS, Seeger JM, Ozaki CK. Directevidence for cytokine involvement in neointimal hyperplasia. Circula-tion 2000;102:1697-1702.

42. Hancock WW, Adams DH, Wyner LR, Sayegh MH, Karnovsky MJ.CD4� mononuclear cells induce cytokine expression, vascular smoothmuscle cell proliferation, and arterial occlusion after endothelial injury.Am J Pathol 1994;145:1008-14.

43. Kolattukudy PE, Quach T, Bergese S, Breckenridge S, Hensley J,Altschuld R, et al. Myocarditis induced by targeted expression of theMCP-1 gene in murine cardiac muscle. Am J Pathol 1998;152:101-11.

44. Martin A, Komada MR, Sane DC. Abnormal angiogenesis in diabetesmellitus. Med Res Rev 2003;23:117-45.

45. Parekh PI, Petro AE, Tiller JM, Feinglos MN, Surwit RS. Reversal ofdiet-induced obesity and diabetes in c57bl/6j mice. Metabolism 1998;47:1089-96.

Submitted Jun 18, 2004; accepted Sep 24, 2004.

Recanalization of arterial thrombus, and inhibition with β-radiation in a new murine carotid...

Documents

Transcript of Recanalization of arterial thrombus, and inhibition with β-radiation in a new murine carotid...