Specific Induction of tie1 Promoter by Disturbed Flow inAtherosclerosis-Prone Vascular Niches and

Flow-Obstructing PathologiesRinnat M. Porat,* Myriam Grunewald,* Anat Globerman, Ahuva Itin, Gregory Barshtein,

Leena Alhonen, Kari Alitalo, Eli Keshet

Abstract—Nonlaminar flow is a major predisposing factor to atherosclerosis. Yet little is known regarding hemodynamicgene regulation in disease-prone areas of the vascular tree in vivo. We have determined spatial patterns of expressionof endothelial cell receptors in the arterial tree and of reporter gene constructs in transgenic animals. In this study weshow that the endothelial cell–specific receptor Tie1 is induced by disturbed flow in atherogenic vascular niches.Specifically, tie1 expression in the adult is upregulated in vascular bifurcations and branching points along the arterialtree. It is often confined to a single ring of endothelial cells functioning as sphincters and hence experiencing the steepestgradient in shear stress. In aortic valves, tie1 is asymmetrically induced only in endothelial cells encountering changesin flow direction. Disturbance of laminar flow by a surgical interposition of a vein into an artery led to induction of tie1,specifically in the region where the differently sized vessels adjoin. In pathological settings, tie1 expression isspecifically induced in areas of disturbed flow because of the emergence of aneurysms and, importantly, in endothelialcells precisely overlying atherosclerotic plaques. Hemodynamic features of atherosclerotic lesion-prone regions,recreated in vitro with the aid of a flow chamber with a built-in step, corroborated an upregulated tie1 promoter activityonly in cells residing where flow separation and recirculation take place. These defined promoter elements might beharnessed for targeting gene expression to atherosclerotic lesions. (Circ Res. 2004;94:394-401.)

Key Words: angiogenesis � atherosclerosis � basic research � blood flow � blood vessels

The luminal surface of blood vessels is constantly exposedto hemodynamic forces, primarily to shear stress that is

the tangential force engendered on endothelial cell surfacesby the blood flow. Hemodynamic forces play a fundamentalremodeling role in the vascular network during its formativestage as well as in adapting the adult vasculature to patho-physiological changes in shear stress (eg, the process ofarteriogenesis1). Hemodynamic forces are also a considerablefactor in the development of vascular pathologies such asatherosclerosis, aneurysms, poststenotic dilatations, and arte-riovenous malformations. Notably, these diseases have thepropensity to develop in vascular regions distinguished bydisturbed flow that occurs naturally in certain vascular niches.Atherosclerosis, although clearly associated with some sys-temic risk factors, is a geometrically focal disease thatpreferentially develops at the outer edges of blood vesselbifurcations and at points of blood flow recirculation andstasis (eg, in aortic valves). In these predisposed locations,fluid shear stress on the vessel wall is significantly lower in

magnitude and exhibits directional changes and flow separa-tion, features absent from regions of the vascular tree gener-ally spared from atherosclerosis.2

The detrimental effects of disturbed flow have inspiredextensive research to elucidate the transduction processes bywhich endothelial cells convert these mechanical stimuli intobiochemical signals and to identify the molecular mediators.These studies, including high-throughput DNA microarrayanalyses, uncovered a large number of shear stress–regulatedgenes that, together, function in multiple regulatory pathwaysin endothelial cells.3,4 These studies, however, were mostlycarried out in cultured endothelial cells, ie, in the absence ofthe normal regulation exerted by matrix components andpossibly also by circulating factors and without the cross-talkwith periendothelial cells. Furthermore, because it is difficultto recreate in vitro the complex flow patterns prevailing inatherogenic microenvironments of the vessel wall, it is notknown which pairwise comparison of shear conditions is ofrelevance in vivo. Although more recent in vitro studies have

Original received July 1, 2003; first resubmission received September 16, 2003; second resubmission received October 31, 2003; revised secondresubmission received November 24, 2003; accepted December 1, 2003.

From the Departments of Molecular Biology (R.M.P., M.G., A.I., E.K.), Surgery (A.G.), and Biochemistry (G.B.), The Hebrew University-HadassahMedical School, Jerusalem, Israel; Virtanen Institute (L.A.), University of Kuopio, Finland; and Molecular Cancer Biology Laboratory (K.A.), Universityof Helsinki, Finland.

*Both authors contributed equally to this study.Correspondence to Dr Eli Keshet, Department of Molecular Biology, The Hebrew University-Hadassah Medical School, Jerusalem 91120, Israel.

E-mail Keshet@cc.huji.ac.il© 2004 American Heart Association, Inc.

Circulation Research is available at http://www.circresaha.org DOI: 10.1161/01.RES.0000111803.92923.D6

394 at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

better approximated the hemodynamic features of atheroscle-rosis-prone regions,5,6 in situ analysis of gene expression inthe native context of atherosclerosis-prone areas has beenlimited. Upregulated gene expression in atherosclerosis-proneareas has been demonstrated for the junction molecule con-nexin43,7 vascular cell adhesion molecule (VCAM)-1,8 andcertain transcription factors.9,10

We have used a candidate gene approach in conjunctionwith in situ hybridization analysis of the native arterial treeand of vessels with flow-obstructing pathologies. To uncoverpromoters induced in predisposed vascular niches, we havealso examined reporter gene constructs in transgenic miceand rats. Endothelial cell–specific receptors are natural can-didates for sensing changes in hemodynamic forces and forparticipating in the mechanotransduction process. Recentstudies showing that shear stress induces a rapid phosphory-lation of endothelial transmembrane receptors and their con-comitant association with adapter proteins in the absence of aligand11–13 have additionally focused our attention on endo-thelial tyrosine kinase receptors.

The subject of this study is the endothelial cell–specific,tyrosine kinase receptor Tie1. Despite intensive efforts, aligand for Tie1 has not been found. Yet Tie1 is indispensablefor maintaining vascular integrity, and its absence results invessel rupture and hemorrhages.14,15 Although Tie1 is abun-dantly expressed in the embryonic vasculature, expressiongrossly subsides concomitantly with vessel maturation. How-ever, previous studies16 and findings reported below haveshown that constitutive expression of tie1 is maintained inselected areas of the adult vasculature, suggesting that Tie1,in addition to its established role in stabilizing embryonicvessels, might also play a role in maintaining proper vascularfunction in the adult.

In this study we show that tie1 expression is selectivelyinduced in areas constantly exposed to disturbed flow, ie, theareas prone to develop atherosclerosis. Moreover, we showthat tie1 is focally upregulated as a consequence of patholog-ical flow obstructions, including by the protrusion of athero-sclerotic lesions into the lumen. These findings, in conjunc-tion with the identification of promoter elements sufficient totarget expression of a surrogate gene to areas of disturbedflow, suggest the potential utility of tie1 promoter in targetingexpression of therapeutic gene products to atheroscleroticlesions.

Materials and MethodsAnimalsTie1-lacZ transgenic mice were previously described.17 Tie-lacZtransgenic rats (outbred Wistar strain HsdBrl:WH) were generatedby the standard pronuclear microinjection technique using the sameconstruct that was used in the production of Tie-lacZ transgenicmice. The zygotes were transferred into the oviducts of the pseudo-pregnant recipient females immediately after microinjection. Thetransgenic line UKUR11 was selected for additional studies and bredto homozygosity. Apolipoprotein (Apo) E–deficient mice18 on aC57BL/6 background were fed Western diet, and 6- to 9-month-oldanimals were monitored for the appearance of aortic lesions.

Vein GraftsEpigastric vein (EV) to common femoral artery (CFA) interpositiongrafts were performed in rats as previously described.19 Briefly, male

Sabra rats (220 to 420 g) were anesthetized, and the CFA and EVwere exposed. The superficial circumflex iliac branch of the CFAwas isolated, and a 1-cm-long segment of the artery was resected. A1-cm-long segment of the ipsilateral EV was harvested, irrigatedwith heparinized saline solution, and installed as a reversed interpo-sition graft. Total ischemic time was �90 minutes, and graft patencywas confirmed by visual inspection. Vein grafts were harvestedalong with the flanking aortic segments at day 67 after operation andfixed in 4% paraformaldehyde. At the time of harvest, graftsremained patent and with no detectable thrombi.

In Situ HybridizationParaffin-embedded aortas and heart caps from ApoE-deficient micewere sectioned, processed, and hybridized in situ as previouslydescribed using 35S-labeled riboprobes.20 Tie1-specific and vascularendothelial growth factor receptor (VEGF-R2)–specific probes wereamplified from the respective cDNAs and cloned into the TOPO-PCRII vector (Invitrogen) using the following primers: Tie1, 5�-CCTGGGCCCTGCCTCACCC-3� and 5�-GGGGGGGCGCT-CATAGGGC-3�; VEGF-R2, 5�-GTGCAGGATGGAGAGCAAGG-3�and 5�-TGGACTCAATGGGCCTTCCATTTCTGTACC-3�.

X-Gal StainingTissues from transgenic rats and mice were fixed in buffered 4%paraformaldehyde. Tissues were then washed three times for 30minutes at room temperature in PBS (pH 7.4) containing 2 mmol/LMgCl2, 0.02% NP40, and 0.01% Na-deoxycholate and stainedovernight at 37°C in the same buffer supplemented with 5 mmol/Lpotassium-ferricyanide, 5 mmol/L potassium-ferrocyanide, and 1mg/mL X-gal. Specimens were then fixed in formalin, embedded inparaffin, sectioned, and counterstained with H&E or, alternatively,examined as whole mounts.

Immunostaining of Whole-MountRetina PreparationsX-gal–stained retinas were washed in PBS and permeabilized in0.5% Triton/PBS for 2 hours before adding �-smooth muscle actinantibody (Sigma) and overnight incubation at 37°C with gentleshaking. After five 1-hour washes in PBS, a secondary anti-mouseantibody conjugated to HRP (Amersham) was added in 0.5%Triton/PBS/1% BSA and retinae-incubated for 3 hours at roomtemperature. Specimens were then stained with 3-amino-9-etylcabazole (Sigma), mounted on glass slides, and viewed in a lightmicroscope.

Flow Apparatus Generating Laminar Flow orDisturbed FlowA parallel plate chamber, designed in the laboratory of Yedgar andcolleagues, as previously described,21 was used. Briefly, a siliconrubber gasket was sandwiched between two plastic plates, creating agap of 250 �m, and a slide covered with test cells was inserted in ahole drilled in the lower plate. The chamber was connected to arecirculating flow circuit composed of a peristaltic pump, a pressuretransducer, and a reservoir with culture medium maintained at 37°C,pH 7.4, and gassed with 5% CO2 and 95% air. Flow velocity wasadjusted to 20 mL/min, generating shear stress of calculated 12dyne/cm2. To generate conditions of flow separation, recirculation,and reattachment, a similar apparatus was designed, except that theslide on top of which cells were seeded contained in its middle a200-�m-high, descending step (see Figure 6A). Confluent cultures ofprimary bovine aortic endothelial cells at a low passage number(�10) on gelatin-coated slides were subjected to flow and subse-quently analyzed as described below.

Infection of Endothelial Cells With apTie-Luciferase Adenovirus and In Situ Mappingof Luciferase ActivityThe same 735-bp-long segment of the tie1 promotor used in thetransgenic experiments (see below) was used to generate a Tie1

Porat et al Induction of Tie1 by Disturbed Flow 395

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

promoter-luciferase adenovirus (also containing a cytomegaloviruspromoter–driven green fluorescent protein by the pAdTrack sys-tem22). For infection, cells were preincubated in the absence ofserum for 1 hour and then infected at a multiplicity of infection of 50to 100 plaque-forming units per cell.

To determine total luciferase activity, cells were scraped off theslide and lysed, and luciferase activity quantified using a dualdetection kit (Promega) according to the manufacturer’s instructions.Redistribution of luciferase activity after flow was determined byincubating the same cell-seeded slide with luciferin for 5 minutes atroom temperature (both before and after flow) and in situ detectionof emitted light in each case with the aid of a cooled CCD camera(Roper chemiluminescence imaging system, model LN/CCD-1300EB), as described by Honigman et al.23 The integrated lightacquired during a short exposure was presented as pseudo-colorimages (blue, least intense; red, most intense), and relative quanti-fication of luciferase activity in successive fields was performedusing the software supplied with the cooled charge-coupled device(CCCD) camera.

An expanded Materials and Methods section can be found in theonline data supplement available at http://www.circresaha.org.

ResultsUpregulation of tie1 Promoter Activity in AorticFlow Dividing PointsReasoning that hemodynamic regulation of tie1 will bereflected in a nonuniform expression along the arterial vas-cular tree, we mapped sites of upregulated tie1 expressionalong the arterial system. To provide a convenient way forspatial resolution, we used tie1 promoter �-galactosidasereporter transgenic mice and rats. The transgenic animalsused contained 735 bp of proximal mouse promoter se-quences cloned upstream of the reporter gene. Previousstudies have shown that this segment of the promoter issufficient to confer a pattern of expression similar to that ofthe endogenous gene.16,17 The arterial tree was isolated fromadult animals and analyzed as a whole-mount preparation for�-galactosidase activity. Strikingly, �-gal activity was mostlyrestricted to arterial segments located at vascular junctions

and downstream of bifurcations and branching points, ie, atsites experiencing shear stress gradients or constantly ex-posed to a disturbed flow (Figure 1). Notably, tie1 is stronglyexpressed at the aorta-renal artery junction, a site distin-guished by its almost straight branching angle and, conse-quently, a site particularly exposed to disturbed flow andexceedingly prone to atherosclerosis (Figure 1, inset).

Upregulated tie1 Expression in MicrovascularBranching Points and Capillary SphinctersAnticipating that induction of tie1 expression at sites exposedto disturbed flow will be manifested also in the microvascu-lature, we analyzed tie1 promoter activity in the vascularnetwork of the retina. To make sure that the observed patternof expression is dictated by the promoter sequences ratherthan being effected by the position of transgene integration,both transgenic rats (Figure 2, top) and transgenic mice(Figure 2, bottom) harboring the same tie1 promoter-�-galactosidase transgene were examined. In both species,tie1 promoter activity was found to be strongly upregulated atpoints of primary and secondary branching from main arte-rioles. It is predominantly upregulated in capillary endothelialcells residing just downstream of the branch point, ie, in cellsexperiencing the steepest gradient in shear stress. Interest-ingly, these endothelial cells are also functionally distin-guished from their neighbors with regard to their increasedassociation with vascular smooth muscle cells (VSMCs),qualifying them as capillary sphincters (see inset in bottomimage). Specific induction of tie1 in capillary sphincters wasnot limited to the retina and was also visible in the micro-vasculature of other tissues, like skin (data not shown).

Asymmetric Expression of tie1 in AorticValve EndotheliumThe aortic and pulmonary artery valves (also known as thesemilunar valves) represent a unique niche within the cardio-

Figure 1. Tie1-lacZ expression in branch-ing point of the arterial tree. X-gal stainingof the femoral artery and its branches iso-lated from a mature tie1-lacZ transgenicrat. Direction of flow is indicated by thered arrow. Black arrows point to flow-dividing sites where upregulated tie1 pro-moter activity is evident. Inset, Aorta-renalartery junction in a 3-month-old tie1-lacZtransgenic mouse. The aorta was cut lon-gitudinally and opened. Arrowheads pointto inlets to other aortic branches.

396 Circulation Research February 20, 2004

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

vascular system with respect to the hemodynamic forcesexerted on the endothelium. Hydraulic models, simulating theaortic valve, have demonstrated flow recirculation and eddycurrents that swirl behind the flexible cusps during rapid flowthrough the valve orifice.24 In the aortic outlet, �-gal activitywas restricted to the valves and was undetectable in the vesselwall downstream (Figure 3A). Sectioning through the outflowtract has shown that tie1 promoter is exclusively expressed inendothelial cells lining the inner aspect of the cup-shapedcusps but not in endothelial cells at the outer aspect of thesame leaflet (Figures 3B and 3C). This unique pattern ofexpression was specific for tie1 and was not detected inVEGF-R2-lacZ mice (data not shown). This asymmetricalpattern of expression strongly argues for hemodynamic reg-

ulation, ie, that flow conditions experienced only by thesecells are conducive for tie1 induction.

Specific Induction of tie1 Expression inAtherosclerotic LesionsTo determine whether flow obstruction induced by luminalprotrusion of atherosclerotic lesions will result in tie1 induc-tion, we carried out an in situ hybridization analysis deter-mining the spatial distribution of endogenous tie1 mRNA inplaque-bearing ApoE-null mice. As shown in Figure 4, tie1mRNA was dramatically upregulated in the endotheliumoverlying atherosclerotic plaques. Remarkably, the area ofupregulated tie1 expression (quantified as 8-fold increase)precisely colocalized with the lesion area and was sharplydemarcated from the nonexpressing flanking healthy endo-thelium. Upregulated tie1 expression was observed already inearly, small, and relatively acellular lesions (Figures 4A and4B) and persisted in advanced plaques (Figures 4C and 4D).

This unique pattern of expression was specific to tie1 andwas not observed with other endothelial cell–specific tyrosinekinase receptors, such as VEGF-R2 (Figure 4F), VEGF-R1,and tie2 (data not shown). A closer inspection of the endo-thelial surface revealed that the tie1-expressing cells, partic-ularly those positioned almost perpendicular to the directionof flow, have rearranged into a tile-like configuration (Figure4E).

Atherosclerotic lesions developed in the aortic sinuses ofApoE-null mouse fed Western diet were also examined for insitu tie1 expression. As shown in Figures 4G and 4H, tie1 wasalso induced in the endothelium overlying lesions at thisparticularly atherogenic site.

Induction of tie1 Expression by Disturbed Flow inAneurysms and Vein GraftsAtherosclerotic lesions represent a complex situation wherehemodynamic regulation might be compounded by the effectof different cytokines produced in the underlying plaque. Wewished, therefore, to examine a situation of disturbed flow inan otherwise healthy endothelium. Abdominal aneurysms

Figure 2. Tie1-lacZ expression in whole-mount retina. X-galstaining of whole-mount retina from mature (21-day-old) tie1-lac-Z transgenic rat (top) and mouse (bottom). Inset, Doublestaining for �-gal activity and �-smooth muscle actin (red)showing extensive coverage with pericytes and VSMCs of tie1-expressing endothelial cells at capillary sphincters.

Figure 3. Tie1-lacZ expression in aorticvalves. A, Whole-mount view of the aorticoutlet of a tie1-lacZ transgenic mousestained for �-gal. Arrow indicates thedirection of blood flow. B, Section of thesemilunar valves cut at a plane perpen-dicular to the direction of flow. C, Highmagnification of the marked area shownin B.

Porat et al Induction of Tie1 by Disturbed Flow 397

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

spontaneously developed in aged tie1-lacZ transgenic ani-mals were examined for tie1 promoter expression. Although�-gal–positive cells were never detected in the same region ofthe healthy aorta, the tie1 promoter was strongly upregulateddownstream of the aneurysmal dilatation (see online Figure1S, available in the online data supplement athttp://www.circresaha.org).

To examine whether tie1 expression is upregulated whenlaminar flow is disturbed by means of a surgical manipula-tion, vein-to-artery interposition grafts were performed asdescribed in Materials and Methods. The grafts were har-vested more than 2 months after surgery to assure that anyinduced trauma had been resolved and possible endothelialcell injuries had been repaired. Reasoning that a complexpattern of disturbed flow would be created near the junctions

where vessels of a markedly different caliber are adjoined, wesubjected the grafts to an in situ hybridization analysis withtie1. As shown in Figure 5, expression of the endogenous tie1gene was indeed induced, preferentially in vein endothelialcells close to the A/V junction. Notably, expression of tie1mRNA was not significantly upregulated in the more distalvein endothelium, suggesting that the pattern of disturbedflow presiding close to the junction is responsible for upregu-lated expression.

Tie1 Is Negatively Regulated by Shear StressIn VitroThe particular sites of upregulated tie1 expression have incommon areas of reduced shear stress or even stasis. There-fore, expression of tie1 was determined in vitro under

Figure 4. In situ hybridization of tie1mRNA in atherosclerotic lesions. Sectionsthrough atherosclerotic lesions developingin the abdominal aorta of 6- to 9-month-old ApoE-null mice continuously fed ahigh-fat diet were hybridized in situ with atie1-specific probe (all images except F),counterstained with H&E, and photo-graphed using bright-field (A, B left, C, E,G) and dark-field (B right, D, F, and H) illu-mination. Green arrows mark the lesionboundaries. High-magnification images likethose shown in A (for an early lesion) andE (representing a 10-fold enlargement ofthe boxed area in the advanced lesionshown in C) enable the relative quantifica-tion of signal intensities by counting auto-radiographic grains in lesion and flankingendothelial cells. Endothelial cells overlyingthe lesion expressed on average 22.4�4.1grains/cell, whereas endothelial cells justupstream or downstream of the lesionexpressed 2.9�1.8 grains/cells and3.8�1.6 grains/cell, respectively (based oncounting �10 cells/region for each lesionin 3 different experiments). F, In situhybridization with a VEGF-R2–specificprobe of a section adjacent to D. Notethat this receptor is not upregulated in thelesion. G and H, Bright-field and dark-fieldimages, respectively, of a section throughan aortic valve bearing several atheroscle-rotic lesions (indicated by arrows), hybrid-ized with tie1-specific riboprobe. Aw indi-cates aortic wall; as, aortic sinus.

398 Circulation Research February 20, 2004

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

conditions of stasis and at different times after the onset oflaminar flow conditions that engender the endothelial cellmonolayer with a shear stress of �12 dyne/cm2. Negativeregulation of tie1 expression by shear stress was indeeddemonstrated for both tie1 mRNA and TIE1 protein as wellas for tie1 promoter activity (see online Figure 2S).

Spatial Mapping of tie1 Promoter Activity in aStep-Flow Chamber Simulating Disturbed FlowTo better approximate the in vivo situation, we aimed torecreate in vitro a gradient of fluid shear stress and other flowphenomenon present in atherosclerosis-prone areas, particu-larly flow separation and recirculation. To this end, a flowchamber apparatus was used in which endothelial cells areseeded on top of a slide with a central backward-facing (ie,descending) step (see Materials and Methods and Figure 6A).Previous studies using a similar device have demonstrated itsutility in modeling major hemodynamic factors present inatherosclerotic lesions.5,6,25 To spatially correlate modula-tions in tie1 promoter activity with graded variations inhemodynamic factors experienced by endothelial cells atdifferent locations, a tie1 promoter-luciferase reporter wasused in conjunction with in situ mapping of luciferaseactivity. Briefly, endothelial cells were preinfected with anadenovirus vector encoding a tie1 promoter-driven luciferase,and a uniform infection of �90% cells was demonstratedwith the aid of a cytomegalovirus promoter–driven greenfluorescent protein reporter contained in the same vector. Aconstant flow was then applied for 16 hours to allowsufficient time for reaching a hemodynamic steady state andfor gene expression reprogramming. The spatial distributionof luciferase activity was determined with the aid of asensitive light-detection CCCD camera and software forproducing pseudo-color images depicting relative light inten-sities in situ (see Materials and Methods). Before the onset offlow, a uniform distribution of luciferase activity was ob-served (Figure 6C). Exposure of the same culture to adisturbed flow, however, led to redistribution of luciferaseactivity, depending on the position of cells relative to thepoint where laminar flow has been disturbed (Figure 6D).Most pronounced upregulated activity was confined to anarrow band of cells located just downstream of the step.

Remarkably, this region fully coincided with the area distin-guished by flow separation and recirculation (Figure 6B), asdeduced from previous studies where streamlines were cal-culated for a similar backward-facing step apparatus. Forexample, in an apparatus with identical geometry (ie, step

Figure 5. In situ hybridization of tie1mRNA in vein interposition grafts. A, Sec-tion through a CFA/EV junction hybridizedwith a tie1-specific riboprobe and coun-terstained with H&E. The contours of theartery and veins are marked by blue lines.Arrowheads point to the junction of theartery and grafted vein, and the arrowpoints to the direction of flow. B, Dark-field image of the same section showingupregulated tie1 mRNA expression imme-diately downstream of the A/V junction(green arrows). C and D, Higher magnifi-cation of the upper and lower areas,respectively, boxed in A. Note that thehybridization signal is in endothelial cells.CFA indicates common femoral artery;EV, epigastric vein; and L, lumen.

Figure 6. In situ mapping of tie1 promoter activity in a flowchamber with a descending step. A, Schematic representationof the flow chamber used with a built-in 200-�m high-descending step. Green blocks indicate the endothelial cellmonolayer, and the arrow indicates the direction of flow. B,Schematic representation of streamlines generated in a similarflow device containing a 200-�m-high backward-facing step.Drawn after Skilbeck et al25 for an illustrative purpose only ofthe zone of flow recirculation beneath the step and a point offlow reattachment a few millimeters downstream (not drawn toscale). C through E, In situ mapping of luciferase activity drivenby a tie1 promoter, determined as described in Materials andMethods. Light emitted because of luciferase activity was cap-tured by the CCCD camera and presented as pseudocolorimages for static (C) and flow (D) conditions. A relative quantifi-cation is provided in image E. Note redistribution of tie1 pro-moter activity by flow, peaking at the area of flow recirculation.

Porat et al Induction of Tie1 by Disturbed Flow 399

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

height of 200 �m), the point of flow reattachment at the righthand end of the vortices was mapped at �300 �m down-stream of the step.25 As evident from the quantification shownin Figure 6E, luciferase activity in the region of disturbedflow was �4-fold higher than in its flanking regions. Nota-bly, the difference in laminar shear stress existing upstreamand downstream of the step had only a slight effect onpromoter activity, thus highlighting the role of hemodynamicfactors associated with flow disturbance rather than a merechange in wall shear stress as the key factor in promoterinduction.

DiscussionThis study represents a comprehensive in vivo analysis of ahemodynamically regulated promoter with respect to itsspecific induction in restricted niches within the arterial treeand in important vascular lesions associated with an ob-structed flow. Findings reported in this study show that tie1 isdramatically and specifically upregulated in areas of dis-turbed flow coinciding with atherogenic vascular niches.Moreover, it is shown that tie1 expression is induced onexperimental and pathological obstruction of blood flow,including in emerging atherosclerotic plaques and in areas ofdisturbed flow generated by the development of aneurysms.

This in vivo approach has several clear advantages overcommonly used approaches to the study of hemodynamicallyregulated promoters using cultured endothelial cells. First, itintegrates regulatory cues exerted by all components of thevessel wall in their natural contexts, as well as systemicinfluences. Second, expression profiles are determined underthe authentic, complex hemodynamic features presiding ineach and every vascular niche, a situation that is difficult torecreate in vitro. It should be pointed out, however, thatalthough this approach precisely maps the native sites ofupregulated expression, it may come short in determining theresponsible flow parameter (eg, distinguishing disturbedshear stress from a shear stress gradient). Third, because of itssuperior spatial resolution, the in situ methodology enablespinpointing at endothelial cells experiencing conditions con-ducive for promoter activation at the resolution level of singlecells. The latter is exemplified in showing that upregulatedtie1 expression is often confined to a single ring of endothe-lial cells, namely, in capillary sphincters. It should be pointedout, however, that it could not be determined whether tie1was induced by the particular flow pattern at sphincters or,alternatively, by the cyclic strain exhibited by VSMCs onsphincter constriction or by the action of VSMC-secretedfactor. Also, it remains to be determined whether signalingthrough the Tie1 receptor plays a role in augmented VSMCrecruitment.

The specific locales of tie1 expression do not reflect organspecificity, do not distinguish arteries from veins, and areindependent of vessel caliber. Instead, the common denomi-nator of locales showing upregulated tie1 expression is somedisturbance of a uniform shear stress. Two vascular config-urations are particularly noticeable for a high degree ofdisturbed flow as well as for being most atherogenic: branch-ing points with a large branching angle (eg, aortic branchingto the kidney) and sites constantly experiencing flow reversal

(eg, aortic valves). These locales were indeed identified as thesites of most-pronounced tie1 induction (Figures 1 and 3,respectively). The most compelling evidence for the claimthat disturbed flow is the trigger for tie1 upregulation wasobtained from surgical manipulation of vein interpositioning,where tie1 mRNA was specifically induced at sites distin-guished by an abrupt change in vessel diameter (Figure 5).

Certain vascular pathologies introduce disturbances in anotherwise laminar flow or aggravate an already existingdisturbance, notably, atherosclerotic plaques and regionaldilatations associated with aneurysms. As reported here, theselesions indeed led to focal induction of tie1 expression.Hemodynamic induction of tie1 may also explain previousresults of its upregulation in arteriovenous malformations.26

Thus, elevated tie1 expression not only marks atherosclero-sis-prone areas but is also an indicator of vascular lesionsassociated with obstructed flow. A similar situation has beenpreviously reported for VCAM-1. Specifically, elevated lev-els of VCAM-1 expression were observed in endothelial cellsin regions of the mouse ascending aorta and arch that arepredisposed to lesion formation, and, furthermore, VCAM-1was also expressed by endothelial cells in early atheroscle-rotic lesions and by intimal cells in more advanced lesions.8

The latter is different from the case of tie1, where expressionremained exclusive to the endothelium overlying the lesion,also in advanced lesions.

The function of Tie1 and the significance of Tie1 signalingin endothelial cell biology are poorly understood, primarilybecause of the failure to find a ligand. Yet a signal transduc-tion pathway for tie1 was recently elucidated.27 It was alsodemonstrated that proteolytic shedding of the putative ligand-binding exodomain of Tie1 generates a truncated receptorstill capable of ligand-independent signaling.28,29 Interest-ingly, a recent study has shown that downregulation of tie1 inendothelial cells exposed to shear stress in vitro is alsoaccompanied by receptor cleavage.30 Thus, precedents for aligand-independent mechanotransduction in other endotheli-al-specific tyrosine kinase receptors11–13 might also apply fortie1 expressed in vascular microenvironments experiencingdisturbed flow.

The notion that signaling through Tie1 might be requiredfor the endothelium to sustain hemodynamic forces is con-sistent with vascular rupture in tie1-null mice and withfindings that Tie1 signaling induces an antiapoptotic re-sponse.27 Consistent with a role for tie1 signaling in therecruitment of periendothelial cells is our observation of acorrelation between upregulated tie1 expression in capillarysphincters and their association with a higher number ofvascular smooth muscle cells (Figure 2).

The hemodynamically regulated expression patterns de-scribed above were generated using a defined segment ofpromoter sequence. Specifically, 735 bp of the tie1 promoterwas sufficient to confer a tight hemodynamic regulation on asurrogate gene both in vivo and in vitro. This is the firsttransgenic study demonstrating hemodynamic regulation dic-tated by a defined regulatory element. A recent in vitro studyhas delineated a negative shear stress response element to a250-bp element within the tie1 promoter.30

400 Circulation Research February 20, 2004

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

Irrespective of the unknown function of Tie1, findingsreported here might have important implications solely basedon the unique features of its promoter. Conceivably, thecritical promoter sequences might be used for targetingexpression of any gene of interest to atherosclerotic plaques.For example, targeting expression of a gene that will promoteplaque passivation might have a significant benefit. Experi-ments examining the feasibility of harnessing the tie1 pro-moter for targeting expression to atherosclerotic plaquesthrough the use of intra-arterially delivered viral vectors areongoing.

AcknowledgmentsThis work was supported by a grant from the Israel ScienceFoundation. We thank Dr H. Giladi for adenovirus vectors, E. Zeirafor in situ luciferase analysis, Dr S. Yedgar for advice on flowdevices, and R. Sinervirta and A. Järvinen for technical help.

References1. van Royen N, Piek JJ, Buschmann I, Hoefer I, Voskuil M, Schaper W.

Stimulation of arteriogenesis: a new concept for the treatment of arterialocclusive disease. Cardiovasc Res. 2001;49:543–553.

2. Malek AM, Alper SL, Izumo S. Hemodynamic shear stress and its role inatherosclerosis. JAMA. 1999;282:2035–2042.

3. Garcia-Cardena G, Comander J, Anderson KR, Blackman BR, GimbroneMA. Biomechanical activation of vascular endothelium as a determinantof its functional phenotype. Proc Natl Acad Sci U S A. 2001;98:4478–4485.

4. Chen BPC, Li Y-S, Zhao Y, Chen K-D, Li S, Lao J, Yuan S, Shyy JY-J,Chien S. DNA microarray analysis of gene expression in endothelial cellsin response to 24-h shear stress. Physiol Genomics. 2001;7:55–63.

5. Nagel T, Resnick N, Forbes Dewey C, Gimbrone M. Vascular endothelialcells respond to spatial gradients in fluid shear stress by enhanced acti-vation of transcription factors. Arterioscler Thromb Vasc Biol. 1998;19:1825–1834.

6. Davies P, Polacek DC, Handen JS, Helmke BP, DePaola N. A spatialapproach to transcriptional profiling: mechanotransduction and the focalorigin of atherosclerosis. Trends Biotechnol. 1999;17:347–351.

7. Gabriels JE, Paul DL. Connexin43 is highly localized to sites of disturbedflow in rat aortic endothelium but connexin37 and connexin40 are moreuniformly distributed. Circ Res. 1998;83:636–643.

8. Iiyama K, Hajra L, Iiyama M, Li H, DiChiara M, Medoff BD, CybulskyMI. Patterns of vascular cell adhesion molecule-1 expression in rabbit andmouse atherosclerotic lesions and sites predisposed to lesion formation.Circ Res. 1999;85:199–207.

9. Hajra L, Evans AI, Chen M, Hyduk SJ, Collins T, Cybulsky MI. TheNF-�B signal transduction pathway in aortic endothelial cells is primedfor activation in regions predisposed to atherosclerotic lesion formation.Proc Natl Acad Sci U S A. 2000;97:9052–9057.

10. McCaffrey TA, Fu C, Du B, Eksinar S, Kent KC, Bush H, Kreiger K,Rosengart T, Cybulsky MI, Silverman ES, Collins T. High-levelexpression of Egr-1 and Egr-1-inducible genes in mouse human athero-sclerosis. J Clin Invest. 2000;105:653–662.

11. Chen K-D, Li Y-S, Kim M, Li S, Yuan S, Chien S, Shyy JY-J. Mech-anotransduction in response to shear stress. J Biol Chem. 1999;274:18393–18400.

12. Shay-Salit A, Shushy M, Wolfovitz E, Breviario F, Dejana E, Resnick N.VEGF receptor 2 and the adherens junction as a mechanical transducer invascular endothelial cells. Proc Natl Acad Sci U S A. 2002;99:9462–9467.

13. Jong LH, Young KG. Shear stress activates tie2 receptor tyrosine kinasein human endothelial cells. Biochem Biophys Res Commun. 2003;304:399–404.

14. Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, Fujiwara Y,Gendron-Maguire M, Gridley T, Wolburg H, Risau W, Qin Y. Distinctroles for the receptor tyrosine kinases Tie-1 and Tie-2 in blood vesselformation. Nature. 1995;376:70–74.

15. Puri MC, Rossant J, Alitalo K, Berenstein A, Partanen J. The receptortyrosine kinase TIE is required for integrity and survival of vascularendothelial cells. EMBO J. 1995;14:5884–5891.

16. Korhonen J, Polvi A, Partanen J, Alitalo K. The tie receptor tyrosinekinase gene: expression during embryonic angiogenesis. Oncogene. 1994;9:395–403.

17. Korhonen J, Lahtinen I, Halmekyto M, Alhonen L, Janne J, Dumont D,Alitalo K. Endothelial-specific gene expression directed by the tie genepromoter in vivo. Blood. 1995;86:1828–1835.

18. Piedrahita JA, Zhang SH, Hagaman JR, Oliver PM, Maeda N. Generationof mice carrying a mutant apolipoprotein E gene inactivated by genetargeting in embryonic stem cells. Proc Natl Acad Sci U S A. 1992;89:4471–4475.

19. Hirsch GM, Karnovsky MJ. Inhibition of vein graft intimal proliferativelesions in the rat by heparin. Am J Pathol. 1991;139:581–587.

20. Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelial growth factorinduced by hypoxia mediates hypoxia-initiated angiogenesis. Nature.1992;395:843–845.

21. Chen S, Barshtein G, Gavish B, Mahler Y, Yedgar S. Monitoring of redblood cell aggregability in a flow-chamber by computerized image anal-ysis. Clin Hemorheol. 1994;14:497–508.

22. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B. Asimplified system for generating recombinant adenoviruses. Proc NatlAcad Sci U S A. 1998;95:2509–2514.

23. Honigman A, Zeira E, Ohana P, Abramovitz R, Tavor E, Bar I, ZilbermanY, Rabinovsky R, Gazit D, Joseph A, Panet A, Shai E, Palmon A, LasterM, Galun E. Imaging transgene expression in live animals. Mol Ther.2001;4:239–249.

24. Stein PD, Walburn FJ, Sabbah HN. Turbulent stresses in the region ofaortic and pulmonary valves. J Biomech Eng. 1982;104:238–244.

25. Skilbeck C, Westwood SM, Walker PG, David T, Nash GB. Dependenceof adhesive behavior of neutrophils on local fluid dynamics in a regionwith a recirculating flow. Biorheology. 2001;38:213–227.

26. Hatva E, Jaakelainen J, Hirvonen H, Alitalo K, Haltia M. Tie endothelialcell-specific receptor tyrosine kinase is upregulated in the vasculature ofarteriovenous malformations. J Neuropathol Exp Neurol. 1996;55:1124–1133.

27. Kontos CD, Cha EH, York JD, Peters KG. The endothelial receptortyrosine kinase Tie1 activates phosphatidylinositol 3-kinase and Akt toinhibit apoptosis. Mol Cell Biol. 2002;22:1704–1713.

28. Yabkowitz R, Meyer S, Yanagihara D, Brankow D, Staley T, Elliot G, HuS, Ratzkin B. Regulation of Tie receptor expression on human endothelialcells by protein kinase C-mediated release of soluble tie. Blood. 1997;90:706–715.

29. McCarthy MJ, Burrows R, Bell SC, Christie G, Bell PR, Brindle NP.Potential roles of metalloprotease mediated ectodomain cleavage in sig-naling by the endothelial tyrosine kinase Tie-1. Lab Invest. 1999;79:889–895.

30. Chen-Konak L, Guetta-Shubin Y, Yahav H, Shay-Shalit A, Zilberman M,Binah O, Resnick N. Transcriptional and post-translation regulation of theTie1 receptor by fluid shear stress changes in vascular endothelial cells.FASEB J . 2003;17:2121–2123.

Porat et al Induction of Tie1 by Disturbed Flow 401

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from

Fig.1S. Induction of tie1 expression due to flow obstruction in aneurysms.

A. A segment of the abdominal aorta of a 2 years old tie1-lacZ transgenic rat with two spontaneously

developed aneurysms stained for b-gal activity. Black arrows mark the boundaries of pathological

vessel dilatations and the broken arrows the direction of blood flow. Note strong induction of tie1 in

the aortic segment flanked by the lesions. B A thin section of the aortic segment shown above. C A

higher magnification of the downstream aneurysmal dilatation. Discontinuity of the elastic lamina,

highlighted by specific staining (not shown), confirmed the boundaries of the lesion as indicated by

the arrows. Note strong b-gal activity upstream of the aneurysm (i.e. in the healthy endothelium

located downstream of the upper lesion). In this section most endothelial cells in the dilation itself

2

were denuded. Note, however, b-gal expression in the remaining endothelial cells in the dilated

region (blue arrowhead) and just downstream of it (black arrowhead).

Fig.2S. Tie1 is negatively regulated by shear stress in vitro.

BAE cells placed in a parallel plate flow chamber were subjected to flow conditions of 12 dyne/cm2

for the indicated times, or maintained in static conditions as controls (S). This particular shear was

chosen because it is higher than shear values experienced by the endothelium in predisposed regions

(estimated to be on the order of 4 dyne/cm2 ) and somewhat lower than shear values existing in

regions generally spared from atherosclerosis (estimated greater than 12 dyne/cm2 ). Cells were then

analyzed for tie1 mRNA, protein and promoter activity as described below

A. Northern blot analysis for tie1 mRNA (arrow). B. Western blot analysis for Tie1 protein. The

glycosylated membrane receptor (upper arrow) and intracellular receptor (lower arrow) are indicated.

3

Note that the significant level of tie1 mRNA and protein expressed by endothelial cells under static

conditions is reduced within a few hours of flow to a barely detectable level. C. Luciferase activity in

BAE cell pre-infected with an adenovirus containing a tie1 promoter-Luc reporter and a CMV

promoter-GFP reporter used for standardization. Luciferase activity in the monolayer subjected to 16

hrs of shear was 6-fold lower than the static control, thereby confirming that upregulation of tie1

expression by a reduced shear stress is at the transcriptional level.

Methods:

Analysis of Tie1 mRNA and protein

For tie1 mRNA analysis, cells were rinsed in PBS, scraped into TRI Reagent (Sigma) and total RNA

was prepared according to the manufacturer’s instructions. 8 mg RNA (obtained from a pool of two

slides for each time point) was electrophoresed through a 1% formaldehyde/ agarose gel, blotted and

hybridized with a 32P-labeled tie1 cDNA probe.

For Tie1protein, cells were scraped, rinsed in ice-cold PBS and lysed in lysis buffer. 50 mg of total

protein was resolved in SDS-PAGE and immunoblotted with an antibody recognizing a carboxy-

terminal peptide of the receptor (Santa-Cruz Biotechnology Inc.). Tie1 protein was then detected

using goat anti-rabbit IgG conjugated to horseradish peroxidase and an ECL detection system

(Amersham).

Alhonen, Kari Alitalo and Eli KeshetRinnat M. Porat, Myriam Grunewald, Anat Globerman, Ahuva Itin, Gregory Barshtein, Leena

Niches and Flow-Obstructing Pathologies Promoter by Disturbed Flow in Atherosclerosis-Prone Vasculartie1Specific Induction of

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 2003 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.0000111803.92923.D62004;94:394-401; originally published online December 11, 2003;Circ Res. 

http://circres.ahajournals.org/content/94/3/394World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circres.ahajournals.org/content/suppl/2004/02/18/94.3.394.DC1.htmlData Supplement (unedited) at:

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Researchin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

at Wyeth Research on August 25, 2015http://circres.ahajournals.org/Downloaded from