Journal of the American College of Cardiology Vol. 52, No. 4, 2008 © 2008 by the American College of Cardiology Foundation ISSN 0735-1097/08/$34.00 Published by Elsevier Inc. doi:10.1016/j.jacc.2008.03.055

PRECLINICAL RESEARCH

Blockade of Angio-Associated Migratory Cell Inhibits Smooth Muscle Cell Migration and Neointima Formation in Accelerated Atherosclerosis

Felix Vogt, MD,* Alma Zernecke, MD,‡ Marie Beckner, PHD,§ Nicole Krott, MSC,*† Anja-Katrin Bosserhoff, PHD,ʈ Rainer Hoffmann, MD,* Marc A. M. J. Zandvoort, PHD,# Thomas Jahnke, MD,¶ Malte Kelm, MD,* Christian Weber, MD,‡ Rüdiger Blindt, MD* Aachen, Regensburg, and Campus Kiel, Germany; Pittsburgh, Pennsylvania; and Maastricht, the Netherlands

Objectives The aim of this study was to elucidate the role of angio-associated migratory cell protein (AAMP) for the migration of vascular smooth muscle cells (SMCs) and for the development of neointimal hyperplasia after vascular injury.

Background Although AAMP has been shown to participate in angiogenesis and cancerogenesis and is predominantly expressed in cells with a migratory phenotype, involvement of AAMP during neointima (NI) formation after arte- rial injury has not been analyzed previously.

Methods The AAMP content in SMCs was examined using 2-photon laser-scanning microscopy and subcellular fractioning. Migratory potential of SMCs transiently transfected with AAMP expression vectors, transfected with small inter- fering ribonucleic acid (siRNA), or treated with antirecombinant angio-associated migratory cell protein– (anti-rAAMP-ab) was examined using transwell migration chamber assays. Expression of AAMP was determined in the atherogenic apolipoprotein E knockout (apoEϪ/Ϫ) mouse model and in the porcine coronary restenosis model by immunohistochemistry and by Western blot. ApoEϪ/Ϫ mice were treated intraperitoneally with anti- rAAMP-ab, and wire-injured carotid arteries were examined.

Results Angio-associated migratory cell protein is localized in the membrane of SMCs, and its expression is enhanced in NI-derived SMCs. The AAMP overexpression increases, while both treatment with anti-rAAMP-ab and transfection with siRNA decreases SMC migration. Knockdown of AAMP decreases RhoA activity in the membrane fraction of SMCs. The AAMP expression by SMCs is enhanced in both animal models. Anti-rAAMP-ab reduces neointimal SMC density at 1 week and NI formation at 4 weeks in apoEϪ/Ϫ mice without affecting proliferation of SMCs.

Conclusions These data reveal an important functional role of AAMP in the migration of SMCs, identifying AAMP as a poten- tial target to limit lesion formation after injury. (J Am Coll Cardiol 2008;52:302–11) © 2008 by the American College of Cardiology Foundation

Angio-associated migratory cell protein (AAMP) is a re- melanoma cells and cells of the colon. cently identified small protein with a molecular weight of 52 Thus, a role in angiogenesis and cancerogenesis was pre- kDa. It is predominantly expressed in cells with a migratory sumed (1,2). Angio-associated migratory cell protein shares phenotype, including endothelial cells, activated T-cells, with immunoglobulin superfamily and glia cells, as well as in malignant tissue such as human members, including cellular adhesion molecules such as neural cell adhesion molecule (NCAM), platelet-endothelial cell

From the *Department of Cardiology, †Interdisciplinary Center of Clinical Research adhesion molecule (PECAM), and leukocyte function (IZKF) “BIOMAT” within the faculty of Medicine, ‡Institute of Molecular Cardio- antigen-2 (LFA-2); these and other immunoglobulin su- vascular Research, RWTH Aachen University, Aachen, Germany; §Department of perfamily mediate adhesion and migration of var- Pathology, Pittsburgh University Medical Center, Pittsburgh, Pennsylvania; ʈInstitute of Pathology, University of Regensburg, Regensburg, Germany; ¶Department of ious malignant circulating cells (1,3). Diagnostic Radiology, University Clinics Schleswig-Holstein, Campus Kiel, Ger- many; and the #Department of Biophysics, Cardiovascular Research Institute Maas- See page 312 tricht, University of Maastricht, Maastricht, the Netherlands. This research project was supported by a grant from the Interdisciplinary Center for Clinical Research “BIOMAT” within the faculty of Medicine at the RWTH Aachen University. The Migration of smooth muscle cells (SMCs) into the TPLSM was financed by the Dutch Scientific Organization (NWO 902-16-276). Manuscript received October 1, 2007; revised manuscript received February 4, developing neointima (NI) is a pivotal step during athero- 2008, accepted March 4, 2008. sclerosis development in native vessels as well as in reste- JACC Vol. 52, No. 4, 2008 Vogt et al. 303 July 22, 2008:302–11 AAMP and Accelerated Atherosclerosis nosis formation after vascular intervention. After injury, a immunohistochemistry. Due to Abbreviations switch of the SMC phenotype from the quiescent to the its inhibitory capacity (12), the and Acronyms secretory phenotype is followed by SMC migration and antibody was used for transwell angio-associated ؍ AAMP production of extracellular matrix in the NI (4,5). During migration chamber assays and migratory cell protein

؍ attachment and migration, SMCs polarize and extend for intraperitoneal administra- anti-rAAMP-ab protrusions in the direction of migration. These microex- tion in mice. Nonimmune rabbit recombinant angio- tensions are driven by actin polymerization, representing a immunoglobulin G serum (Sigma- associated migratory cell dynamic reorganization of the actin cytoskeleton. The regula- Aldrich, St. Louis, Missouri) protein–antibody apolipoprotein ؍ tory mechanisms of actin restructuring remain unclear, but a served as negative control. For apoE؊/؊ complex interaction of actin-cytoskeletal associated proteins detection, an alkaline phospha- E knockout fluorescein ؍ with GTPases might contribute (6). These processes possibly tase coupled secondary antibody FITC isothiocyanate induce a change from a quiescent to a migratory SMC (anti-rabbit, Chemicon, Temecula, green fluorescent ؍ phenotype (7,8). Various mediators such as interleukins or California) was used. Antibody GFP protein platelet-derived growth factor—which can be secreted from staining by monoclonal antibod- human smooth ؍ hSMC platelets, monocytes, or SMCs—can trigger the migration of ies against alpha-smooth muscle muscle cell SMCs (9). The expression of AAMP by SMCs during actin (SMA) (Dako, Glostrup, neointima ؍ NI atherosclerosis and restenosis development, however, has not Denmark), beta-actin (Sigma- ؍ been studied previously. Thus, the rationale of this study was to Aldrich), von Willebrand factor rSMC rat smooth muscle cell elucidate the role of AAMP for the migration of SMCs and (Dako), RhoA (Santa Cruz Bio- small interfering ؍ siRNA for the development of neointimal hyperplasia after vascular technology, Santa Cruz, Califor- ribonucleic acid injury. To explore this, we quantified AAMP expression in nia), and Ki67 (Dako) was de- smooth muscle ؍ SMA SMCs derived from the NI and media as well as in athero- tected by a secondary horseradish Ϫ Ϫ actin sclerotic lesions of apolipoprotein E knockout (apoE / ) mice peroxidase-labeled antibody from and restenotic lesions of porcine coronary arteries. Further- Dako. For AAMP fluorescence more, the subcellular localization of AAMP and its effects on staining, a fluorescein isothiocyanate (FITC)-conjugated SMC migration and proliferation were analyzed. Eventually, antirabbit antibody was used (Chemicon). we evaluated the in vivo expression of AAMP in the non- Fluorescence staining, 2-photon laser-scanning micros- atherogenic porcine model of coronary balloon dilation and in copy, and flow cytometric analysis. The rSMCs and the atherogenic mouse model of accelerated lesion formation. hSMCs grown on glass slides were washed in phosphate The effect of AAMP blockade on NI formation as well as buffered saline (PBS), fixed immediately with 3.7% formal- SMC migration and proliferation was tested by treatment of dehyde/PBS for 5 min, dehydrated by immersion in ace- Ϫ Ϫ wire-injured apoE / mice with an inhibitory antibody against tone, and permeabilized with 0.1% Triton X-100. Cells AAMP. were stained with anti-rAAMP-ab and the respective FITC-conjugated antibody. After nuclear staining with DAPI (4=,6-diamidino-2-phenylindole, Sigma-Aldrich), Methods slides were analyzed by 2-photon laser-scanning microscopy Culture of SMCs. Medial rat smooth muscle cells using a Nikon Eclipse E600FN upright microscope (Tokyo, (rSMCs) were isolated from the medial layer of the thoracic Japan), incorporated in the Bio-Rad Radiance 2100MP aorta of 6-week-old male Sprague-Dawley rats; neointimal imaging system, and operated by Lasersharp2000 V6.0 rSMCs were derived from the neointimal thickening of the (Bio-Rad, Hemel Hempstead, United Kingdom). Excita- aorta 2 weeks after balloon angioplasty by microscopic tion was by the Tsunami Ti:sapphire laser (Spectra-Physics, dissection (10). Human smooth muscle cells (hSMCs) were Mountain View, California). Slides were observed through isolated from the medial layer of mammary arteries after a water dipping 60ϫ fluor objective with a 1.00 numerical coronary artery bypass operation as described (11); informed aperture (Nikon, Tokyo, Japan) (13). written consent of all patients was obtained before the Flow cytometric analysis was performed as previously procedure. The SMCs were sparsely and densely grown and described (14). Cells were detached with accutase and consecutively tested for their migratory potential with a subjected to flow cytometric analysis with or without transwell migration chamber system as previously described Tween-induced membrane permeabilization and stained (11). Only SMCs with a distinct migratory phenotype were with anti-rAAMP-ab and the respective FITC-conjugated used for migration experiments; dense-grown SMCs, char- antibody. An FITC-conjugated immunoglobulin G isotype acterized by a significantly lower migratory potential, served served as negative control. as negative control. Quantitative Western blotting. For Western blotting . An affinity-purified polyclonal antibody of cultured SMCs, cells were suspended in radio- against recombinant AAMP was generated in rabbits immunoprecipitation assay buffer (Roche, Mannheim, Ger- (anti-rAAMP-ab; concentration 1 ␮g/ml) (1) and was many); for analysis of SMC supernatants, concentration of used for detection of AAMP by Western blotting and supernatants was performed by freeze drying (Christ Alpha 304 Vogt et al. JACC Vol. 52, No. 4, 2008 AAMP and Accelerated Atherosclerosis July 22, 2008:302–11

2-4, Braun Biotech Int., Melsungen, Germany) before Animal models. All animal study protocols were approved Western blotting. For in vivo expression analysis, porcine by the local Institutional Animal Care and Use Committee coronary arteries were prepared as previously described and conformed to the tenets of the American Heart Asso- (15,16). Always, 15 ␮g of protein was used. Equal loading ciation on research animal use. of Western blots was assured by quantification of total Atherogenic mouse model of accelerated atherosclerosis. protein content before each experiment (BCA Protein An atherogenic mouse model of accelerated lesion forma- Assay Kit, Pierce, Rockford, Illinois) and by assessment of tion was applied as previously described (18). In brief, a beta-actin expression. Blotted membranes were quantified 0.014-inch flexible angioplasty guide wire was advanced Ϫ/Ϫ with “ImageJ” image processing software (17). into the carotid artery of apoE mice (C57BL/6 back- Subcellular fractioning and RhoA and Rac activity as- ground, M&B, Ry, Denmark) via an incision in the external says. To explore differential AAMP and RhoA localization carotid artery, and endothelial denudation was achieved by by subcellular fractioning, a proteome extraction kit (Merck 3 rotary passes along the common carotid artery. Chemicals, Nottingham, United Kingdom) was applied as After intraperitoneal administration of anti-rAAMP-ab described by the manufacturer. Membrane fractions that at different concentrations, serum concentration of anti- contain activated guanosine triphosphate-bound RhoA rAAMP-ab was determined by dot blot reactions at 24, 48, and cytosolic fractions that contain inactive guanosine and 72 h. At 72 h, an anti-rAAMP-ab serum concentration of 2 ␮g/ml was achieved after application of 50 ␮g anti- diphosphate-bound RhoA were analyzed by Western blot- Ϫ/Ϫ ting. For Rac analysis, a Rac activation assay kit (Upstate, rAAMP-ab/animal. Subsequently, apoE mice were in- traperitoneally treated with 50 ␮g anti-AAMP-ab or con- Lake Placid, New York) was applied according to the ϭ manufacturer’s instructions. trol nonimmune serum (n 9 each) every 72 h starting 1 week before and up to 4 weeks after injury. One week Computed AAMP sequence analysis. Searching for po- ϭ ϭ tential transmembrane domains was performed by com- (n 3) or 4 weeks (n 6) after injury, the animals were puted AAMP sequence analysis (DNAMAN software, killed and the arteries were perfusion-fixed with 4% para- formaldehyde and embedded in paraffin. Lynnon BioSoft, Quebec City, Canada). Porcine model of balloon-induced coronary resteno- Transient transfection of AAMP vector constructs and sis. A porcine restenosis model after coronary arterial injury small interfering ribonucleic acid (siRNA). The AAMP was performed as previously described (15,16). Briefly, after complementary deoxyribonucleic acid was inserted into the establishing arterial access by Doppler guided puncture of vector pCMX-GL1 via EcoRI restriction sites. Mock and the femoral artery, 7,500 IU heparin were administered and green fluorescent protein (GFP) vectors were used as the left anterior descending coronary artery was engaged controls. To induce AAMP knockdown in SMCs, siRNA with a 6-F guiding catheter. After intracoronary adminis- was used. The AAMP siRNA, lamin controls, and fluo tration of nitrate (0.1 mg), the vessel diameter was calcu- duplex transfection efficacy controls were purchased from lated online by quantitative coronary angiography. The Dharmacon (Lafayette, Colorado). balloon/artery ratio was 1.1 to 1.2. The noninstrumented ϫ 5 For transfection of AAMP vectors and siRNA, 3 10 left circumflex artery served as uninjured control. Angio- cells were seeded into T25 plates for functional assays or plasty was performed in a total of 6 animals by balloon ϫ 4 5 10 cells were seeded into each well of a 6-well plate for inflation for 2 ϫ 30 s (10 to 12 atm). Animals were protein assays. The Amaxa NucleofectorTM technology maintained on a normal laboratory diet until being killed at (Amaxa, Cologne, Germany) was used for transfection of 1 week (n ϭ 6) or 4 weeks (n ϭ 6). Hearts were harvested, SMCs as previously described (11). and the epicardial coronary arteries were removed. Chemotaxis assays and proliferation analysis. For che- Histomorphometrical and histopathological evaluation. motaxis experiments, transwell migration chambers were For histomorphometrical evaluation of murine arteries, used (Neuro Probe Inc., Gaithersburg, Maryland). Polycar- modified Movat’s pentachrome stainings were performed on ␮ bonate filters (13 mm diameter, 8 m pore size; Whatman paraffin-embedded sections as previously described (17). For International Ltd., Kent, United Kingdom) were coated histopathological evaluation, cell nuclei were counterstained with gelatine (5 mg/ml) as previously described (11). The with Mayer’s hemalaun (Merck, Darmstadt, Germany). upper compartment was filled with Dulbecco’s modified The AAMP expression intensity in SMCs was quantified Eagle’s medium with or without anti-rAAMP-ab (100 by pixel analysis of AAMP staining intensity (Vectastain 5 ␮g/ml) (15). The rSMCs (2 ϫ 10 cells/ml) were placed in ABC-AP, Vector Laboratories, Burlingame, California). the upper compartment of the chamber and incubated for Results were expressed as cumulative intensity divided by ϩ ϩ 4 h at 37°C. Transmigrated cells adhering to the lower total SMC area. Total AAMP or ␣-SMA vessel wall chamber surface were fixed, stained, and counted (Leica area of the media and NI was quantified and expressed as DM RX microscope, Wetzlar, Germany). positive area percentage. For proliferation analysis, a 5-bromodeoxyuridine (BrdU) For proliferation analyses, Ki67-positive cell nuclei were cell proliferation assay from Chemicon was applied accord- counted, divided by total cell numbers, and displayed as ing to the manufacturer’s instructions. percentage index of cell proliferation. Tissue sections from JACC Vol. 52, No. 4, 2008 Vogt et al. 305 July 22, 2008:302–11 AAMP and Accelerated Atherosclerosis the proximal, mid, and distal part of the injured vessel segment were analyzed as previously described (15–17). A Tissue analysis was performed on digitized images (Leica DM RX microscope; “Diskus” analysis software, version 4.30, Hilgers, Koenigswinter, Germany) (15); for pixel analysis, “ImageJ” image processing software was applied. Statistical analysis. All results are expressed as mean Ϯ SEM. For in vitro experiments, statistical significance was evaluated with the unpaired Student t test or analysis of variance, followed by Dunnett’s post hoc test for more than 2 means. For histomorphometric measurements, the 20 µm 20 µm Kruskal-Wallis test was used to determine overall statistical 60 significance, followed by the Mann-Whitney U test with B Bonferroni correction for subsequent pairwise comparison if 50 the overall significance was Ͻ0.05; p values Ͻ0.05 were 40 * P<0.01 considered significant. ~ P<0.001 30 *

20 Results 10

Subcellular localization of AAMP. In sparsely grown AAMP expression [intensity/area] ~ 0 hSMCs, AAMP expression was enhanced and preferen- cytosolic membrane nuclear tially localized along the cell membrane, whereas only weak fraction fraction fraction staining for AAMP was detected in dense hSMCs and AAMP medial rSMCs (Fig. 1A), implying a phenotype-related change in localization and strength of expression. Subcel- lular fractioning of hSMCs confirmed that AAMP is C 4000 located primarily at the cellular membrane (55.5 Ϯ 4.4 3000 * P<0.01 intensity/area) but also in the cytosolic fraction (23.1 Ϯ 3.9; * p Ͻ 0.01), whereas the nuclear fraction did not contain 2000

AAMP (0.2 Ϯ 0.1; p Ͻ 0.001) (Fig. 1B). Flow cytometric cell count 1000

analysis of hSMCs also demonstrated that AAMP is mean FITC-A intensity presented preferentially at the cell membrane (without 0 10 100 1000 10000 nonimmune Ϯ Ø Tween + Tween Tween treatment: 1,906 185 FITC-A intensity) but FITC-A serum can also be found in the cytosol (permeabilized cells with Ϯ Tween treatment: 3,967 203; nonimmune control: 345 D 2.0 Ϯ Ͻ 1.8 25; p 0.01) (Fig. 1C). Accordingly, computed analysis 322 - 345 24aa of AAMP sequence revealed a potential transmembrane do- 1.6 main between amino acids 322 and 345 (Fig. 1D). 1.4 1.2 AAMP expression by SMCs. In medial sparse-grown 1.0 hSMCs, AAMP expression was significantly increased 0.8 compared with dense-grown medial hSMCs (33.3 Ϯ 4.0 hydrophoby score 0.6 and 11.2 Ϯ 3.3 intensity/area, p Ͻ 0.01). Similarly, sparse 0.4 medial rSMCs displayed significantly higher AAMP levels 0.2 0.0 compared with dense medial rSMCs (64.9 Ϯ 6.0 and 21.4 1 109 218 326 434 Ϯ 4.1, p Ͻ 0.01). Both, sparse and dense rSMCs derived Position from the NI displayed a pronounced AAMP expression (72.5 Ϯ 6.1 and 66.2 Ϯ 7.1, p ϭ 0.54), indicating that Figure 1 Expression and Subcellular Localization of AAMP

AAMP expression in neointimal cells can not be downregu- (A) Two-photon laser-scanning microscopy of sparse smooth muscle cells (left) detect- lated (Fig. 2A). ing intensified angio-associated migratory cell protein (AAMP) distribution (green) at the In cellular supernatants of rSMCs as well as hSMCs, cellular membrane (arrows), whereas dense cells (right) contain less AAMP without distinctive distribution patterns. (B) Subcellular fractioning: the cellular membrane con- AAMP expression was very low or not detectable (data not tains Ͼ50% of total cellular AAMP; as expected by fluorescence microscopy, the cellu- shown). lar nuclei did not contain AAMP. (C) Flow cytometric analysis of AAMP expression Effect of AAMP on SMC migration and proliferation. treated or untreated with lysis buffer confirmed the results of the subcellular fraction- ing experiments (red: nonimmune serum, green: without Tween, blue: with Tween). (D) Treatment with the inhibitory anti-rAAMP-ab reduced Computed AAMP protein sequence analysis revealed a potential transmembrane SMC migration. Migration of both medial and, even more domain between amino acids 322 and 345. FITC ϭ fluorescein isothiocyanate. profoundly, neointimal rSMCs was inhibited by AAMP 306 Vogt et al. JACC Vol. 52, No. 4, 2008 AAMP and Accelerated Atherosclerosis July 22, 2008:302–11

Figure 2 Phenotype-Related AAMP Expression in SMCs and the Effect of AAMP on SMC Migration

(A) Western blot analysis of angio-associated migratory cell protein (AAMP) expression in human smooth muscle cells (hSMCs) and rat smooth muscle cells (rSMCs). (B) Blockade of AAMP by an inhibitory recombinant AAMP-antibody (anti-rAAMP-ab) reduced the migratory potential of medial and of neointimal rSMCs in modified trans- well chambers. (C) The migratory activity of hSMCs and rSMCs after AAMP-sense vector-transfection was increased compared with controls. (D) The AAMP knockdown by small interfering ribonucleic acid (siRNA) alleviated hSMC migration compared with controls. (E) Five-bromodeoxyuridine (BrdU) assays showed no effect of anti-rAAMP-ab treatment, vector transfection, and siRNA treatment on SMC proliferation. GFP ϭ green fluorescent protein. JACC Vol. 52, No. 4, 2008 Vogt et al. 307 July 22, 2008:302–11 AAMP and Accelerated Atherosclerosis blockade in comparison with control (21.5 Ϯ 4.0 and 74.4 Ϯ 11.6 cells/area vs. 49.7 Ϯ 3.8 and 164.5 Ϯ 20.6 cells/area, respectively; p Ͻ 0.001) (Fig. 2B). After transfection with AAMP expression vectors, SMCs displayed an enhanced migratory potential compared with mock or GFP trans- fected cells (hSMCs: 48.5 Ϯ 3.5, 26.3 Ϯ 2.7, and 27.3 Ϯ 3.0; p Ͻ 0.01; rSMCs: 55.9 Ϯ 4.2, 31.2 Ϯ 2.4, and 34.4 Ϯ 2.9; p Ͻ 0.01) (Fig. 2C). The AAMP vector transfection resulted in AAMP overexpression as compared with mock and GFP transfected cells (hSMCs: 34.1 Ϯ 2.0, 13.2 Ϯ 2.1, and 16.0 Ϯ 2.8 intensity/area; rSMCs: 41.1 Ϯ 2.0, 18.0 Ϯ 3.0, and 17.5 Ϯ 2.8 intensity/area; p Ͻ 0.01) (Fig. 2C). Accordingly, AAMP knockdown by siRNA reduced the migratory potential of hSMCs compared with untreated and lamin controls (24.9 Ϯ 2.0, 57.9 Ϯ 3.5, and 55.6 Ϯ 3.5 cells/area; p Ͻ 0.01) (Fig. 2D), and the cellular AAMP content was reduced compared with controls (13.1 Ϯ 2.2, 33.3 Ϯ 2.8, and 31.3 Ϯ 3.5 intensity/area; p Ͻ 0.01) (Fig. 2D). Fluo duplex transfection efficacy checks confirmed good transfection efficacy (data not shown). Treatment with anti-rAAMP-ab (0.094 Ϯ 0.003 absor- bance), AAMP sense-vectors (0.090 Ϯ 0.007), and AAMP siRNA (0.096 Ϯ 0.007) did not alter hSMC proliferation compared with controls (untreated: 0.098 Ϯ 0.005; nonim- mune serum: 0.099 Ϯ 0.005; mock transfection: 0.095 Ϯ 0.008; GFP transfection: 0.087 Ϯ 0.002; lamin: 0.094 Ϯ 0.005; p ϭ 0.738) (Fig. 2E). These data imply that AAMP is crucially involved in the regulation of SMC migratory activity. Cellular membrane-associated AAMP regulates SMC migration by modulation of RhoA activity. By proteome extraction, the effect of AAMP knockdown by siRNA was assessed for membrane and cytosolic hSMC fractions. The AAMP knockdown effectively reduced the membranous AAMP content (11.3 Ϯ 1.5 intensity/area) compared with Figure 3 RhoA Activity Is Reduced by AAMP Blockade Ϯ Ϯ Ͻ untreated (24.3 2.0) and lamin controls (23.3 2.7; p (A) The siRNA knockdown reduced AAMP content in the membrane fraction but 0.01) (Fig. 3A). In the cytosolic fraction, AAMP content not in the cytosolic fraction. (B) The RhoA activity assay: in the membrane frac- was not significantly altered by siRNA (14.2 Ϯ 2.5) com- tion, siRNA and anti-rAAMP-ab treatment greatly reduced RhoA activity, whereas Ϯ activated RhoA expression was elevated after AAMP overexpression. In the pared with untreated (18.5 2.0) and lamin controls (19.2 cytosolic fraction, expression of inactive RhoA was not affected by anti- Ϯ 1.8; p Ͼ 0.05) (Fig. 3A). The AAMP knockdown (0.5 Ϯ rAAMP-ab treatment, vector transfection, or AAMP siRNA treatment. Abbrevia- 0.1) and anti-rAAMP-ab treatment (0.4 Ϯ 0.0) both tions as in Figure 2. reduced RhoA content, whereas AAMP sense-vector trans- fection (9.1 Ϯ 1.0) enhanced RhoA content in the mem- brane compared with controls (untreated: 4.6 Ϯ 1.2, mock: of AAMP in 1- and 4-week groups in the media compared Ϯ Ϯ 4.2 Ϯ 1.0, GFP: 5.4 Ϯ 0.9, lamin: 5.1 Ϯ 0.8; p Ͻ 0.01) with the uninjured group (25.3 2.8, 23.1 2.2, and 18.0 Ϯ ϭ (Fig. 3B). Cytosolic RhoA content was not altered by the 1.4 intensity/cellular area, p 0.18) (Figs. 4A and 4B), different treatment modalities (AAMP siRNA: 11.0 Ϯ 1.0; whereas neointimal AAMP expression was significantly anti-rAAMP-ab: 11.3 Ϯ 1.0; AAMP sense-vector: 10.2 Ϯ increased at 1 and 4 weeks as compared with the media of 1.0; untreated: 11.3 Ϯ 0.9; mock: 11.3 Ϯ 0.8 intensity/area; the same group (40.2 Ϯ 2.8 and 36.8 Ϯ 2.6, p Ͻ 0.01) (Figs. GFP: 10.2 Ϯ 0.9; lamin: 11.0 Ϯ 1.0; p Ͻ 0.01) (Fig. 3B). 4A and 4B). Analysis of Rac activation did not reveal significant differ- Similar AAMP expression patterns were observed in ences (data not shown). the porcine model of balloon-induced coronary restenosis Analysis of temporal and quantitative AAMP expression (Fig. 5A). Again, pixel analysis revealed a substantial in vivo. Pixel analysis of immunofluorescence staining in increase in AAMP expression in the NI at 1 and 4 weeks Ϫ Ϫ injured apoE / mice by calculating AAMP expression after injury (39.1 Ϯ 3.1 and 35.0 Ϯ 2.5, p Ͻ 0.01), whereas intensity/cellular area revealed a slightly stronger expression medial expression of AAMP was slightly enhanced at 1 and 308 Vogt et al. JACC Vol. 52, No. 4, 2008 AAMP and Accelerated Atherosclerosis July 22, 2008:302–11

A AAMP nonimmune serum  AAMP nonimmune serum uninjured uninjured uninjured uninjured

100 µm 100 µm at 1 week at 1 week 100 µm 100 µm at 1 week at 1 week

100 µm 100 µm at 4 weeks at 4 weeks

100 µm 100 µm at 4 weeks at 4 weeks

100 µm 100 µm

100 µm 100 µm

B

uninjured at 1 week at 4 weeks AAMP expression [intensity/cellular area] AAMP expression [intensity/cellular 

50 * P<0.01 * 40 Figure 4 Increased AAMP Expression in Murine Carotid Arteries After Wire-Induced Injury 30 * (A) Photomicrographs of angio-associated migratory cell protein (AAMP)– stained and nonimmune serum control murine sections. (B) Respective pixel 20 intensity analysis. NI ϭ neointima. 10

Ϯ Ϯ AAMP expression [intensity/area] 4 weeks compared with control (24.5 2.6, 21.9 1.9, and 0 Ϯ ϭ 17.5 1.1, p 0.22) (Fig. 5B). Sections stained with uninjured at 1 week at 4 weeks nonimmune serum showed only faint nonspecific back- ground staining. Additionally, vessel wall protein lysates of injured and non-injured arteries were analyzed by Western AAMP blotting and quantified by subsequent densitometry. A significantly higher AAMP protein content was demon- strated at 1 and 4 weeks (41.5 Ϯ 2.2 and 22.2 Ϯ 1.2 β-actin intensity/area, respectively) after balloon dilation compared with control vessels (2.8 Ϯ 0.6, p Ͻ 0.01) (Fig. 5C). ␣ Figure 5 Increased AAMP Expression Analysis of AAMP/ -SMA-colocalization. An AAMP/ in Balloon-Injured Porcine Coronary Arteries ␣-SMA-colocalization analysis was performed to examine how SMCs contribute to AAMP expression. (A) Photomicrographs of angio-associated migratory cell protein (AAMP)–stained and ␣ nonimmune serum control porcine sections. (B) Respective pixel intensity analysis. In the murine model, medial AAMP/ -SMA double- (C) Vessel wall protein lysate analysis of AAMP-expression. NI ϭ neointima. positive areas as percentage of the total medial area were JACC Vol. 52, No. 4, 2008 Vogt et al. 309 July 22, 2008:302–11 AAMP and Accelerated Atherosclerosis similar for AAMP and ␣-SMA in uninjured (44.0 Ϯ 3.6% vs. 44.4 Ϯ 2.9%, p ϭ 0.87) and in injured specimen (at 1 A AAMP α-SMA week: 42.0 Ϯ 3.9% vs. 39.6 Ϯ 2.7%, p ϭ 0.87; at 4 weeks: 37.0 Ϯ 3.0% vs. 36.0 Ϯ 3.2%, p ϭ 0.87). Also, the uninjured uninjured neointimal double-positive percentage area was similar within the groups (at 1 week: 32.0 Ϯ 3.4% vs. 33.0 Ϯ 3.2%, p ϭ 0.82; at 4 weeks: 29.7 Ϯ 3.8% vs. 28.3 Ϯ 3.2%, p ϭ 0.82) (Figs. 6A and 6B); due to reduced neotintimal cell density, absolute values for AAMP/␣-SMA double-positive percentage area were lower compared with the media. 100 µm 100 µm at 1 week AAMP contributes to NI formation and SMC migration at 1 week ؊ ؊ but not to proliferation after injury in apoE / mice. Ϫ Ϫ The apoE / mice were treated with inhibitory anti- rAAMP-ab or nonimmune serum intraperitoneally. At 4 weeks after injury, neointimal hyperplasia of the carotid artery was significantly reduced in anti-rAAMP-ab–treated mice compared with nonimmune serum-treated control 50 µm 100 µm 50 µm 100 µm mice (42,100 Ϯ 2,800 and 78,700 Ϯ 3,100 ␮m2,pϽ 0.01), whereas medial areas were not different (19,237 Ϯ 3,572 and 18,423 Ϯ 815 ␮m2,pϭ 0.86) (Figs. 7A and 7B). The B reduction in neointimal area was associated with a signifi- ϩ cant reduction in the relative density of ␣-SMA cells in the NI compared with nonimmune serum-treated mice at 1 week (23.5 Ϯ 3.2% and 39.6 Ϯ 2.7%, p Ͻ 0.01) but not at 4 weeks (37.8 Ϯ 3.9% and 36.0 Ϯ 3.2%, p ϭ 0.73) after injury (Figs. 7C and 7D). Neointimal cell proliferation, as quantified by Ki67 staining, was similar in both groups at 1 (anti-rAAMP-ab: 12.2 Ϯ 0.9%; uninjured: 12.9 Ϯ 1.0%, p ϭ 0.92) and at 4 weeks (anti-rAAMP-ab: 4.2 Ϯ 0.9%; uninjured: 5.1 Ϯ 1.0%; p ϭ 0.57) (Fig. 7E). In the media of uninjured and injured arteries, there was only minimal Ki67 staining of about 1 percentage point (data not shown). These data imply that neutralization of AAMP protects from NI remodeling by early reduction of SMC migration into the developing lesion.

Figure 6 Colocalization of AAMP and ␣-SMA Discussion (A) Colocalization of AAMP and ␣-smooth muscle actin (SMA) in wire-injured Initially, the expression of AAMP was described for a variety sections of the murine model. (B) Quantification of AAMP- and ␣-SMA-positive of cell types, including endothelial cells, activated T-cells, and areas in the murine media and NI, respectively. Abbreviations as in Figure 4. malignant cells, which are all characterized by a migratory phenotype (12). In further studies, upregulation of AAMP was observed during cancerogenesis in gastrointestinal stromal promoted SMC migration, and blockade of AAMP by an tumors (19) and ductal in situ of the breast (20). On inhibitory antibody or by siRNA reduced the migratory activity matrigel, endothelial tube formation and cell migration could of vascular SMCs without affecting proliferation in vitro and in be blocked by an inhibitory AAMP antibody (12). Thus, it was vivo. The study elucidated that, in line with these observations, speculated that AAMP was involved in tumor formation by blockade of AAMP results in a highly significant reduction of Ϫ/Ϫ supporting tumor angiogenesis (21). Although cell migration is NI development in an apoE mouse injury model. Con- also integral during the development of accelerated atheroscle- comitantly, the blockade of AAMP generated an early sub- rosis and restenosis (4), data investigating the role of AAMP stantial decrease of SMC density in the NI. are not available. In summary, these data strongly suggest that AAMP plays In the present study, a significant upregulation of AAMP an important role for SMC migration during the development expression in vascular SMCs derived from the NI and in and progression of accelerated atherosclerosis and restenosis. SMCs from different atherosclerotic and restenotic animal Notably, this was conclusively demonstrated for the non- models was demonstrated. Moreover, AAMP overexpression atherogenic porcine restenotic model and the atherogenic 310 Vogt et al. JACC Vol. 52, No. 4, 2008 AAMP and Accelerated Atherosclerosis July 22, 2008:302–11

Figure 7 AAMP Blockade Reduces Wire-Induced NI Formation and SMC Density in ApoE؊/؊ Mice

(A) Mice intraperitoneally treated with nonimmune serum developed normal neointimal hyperplasia, whereas in the anti-rAAMP-ab–treated group, neointimal area was significantly reduced. (B) Histomorphometrical assessment revealed a marked reduction of neointimal hyperplasia in the anti-rAAMP-ab–treated group, whereas medial areas did not differ. (C and D) One week after anti-rAAMP-ab treatment, neointimal SMC density represented by ␣-SMA staining was reduced compared with control. After 4 weeks, anti-rAAMP-ab treatment did not reduce neointimal SMC density compared with nonimmune serum. (E) Vascular cell proliferation analysis revealed similar pro- liferation indexes (percent Ki67-positive nuclei) in the neointima (NI) at 1 and 4 weeks after injury in both groups. Abbreviations as in Figures 2 and 6. mouse model of accelerated lesion formation. This is of mechanisms of applied mechanical injury vary significantly importance, because both models are characterized by different between rodents and the situation in humans. strengths and weaknesses. The pathophysiologic mechanisms Although prior studies supported a relevance of AAMP for of neointimal development in the porcine coronary model are cellular migration in angiogenesis and cancerogenesis, only most similar to restenotic lesions in humans, but the model has limited data regarding the mechanistic role of AAMP are a nonatherogenic background. The atherogenic mouse model available. Here, it could be demonstrated that AAMP is overcomes this limitation, but the character of the restenotic localized on the cellular membrane and in the cytosol of SMCs lesions in the smaller mouse carotid arteries as well as the and that the membrane-associated pool of AAMP seems to be JACC Vol. 52, No. 4, 2008 Vogt et al. 311 July 22, 2008:302–11 AAMP and Accelerated Atherosclerosis

REFERENCES

1. Beckner ME, Krutzsch HC, Stracke ML, et al. Identification of a new immunoglobulin superfamily protein expressed in blood vessels with a heparin-binding consensus sequence. Cancer Res 1995;55:2140–9. 2. Beckner ME, Liotta LA. AAMP, a conserved protein with immuno- globulin and WD40 domains, regulates endothelial tube formation in vitro. Lab Invest 1996;75:97–107. 3. Beckner ME, Krutzsch HC, Klipstein S, et al. AAMP, a newly identified protein, shares a common epitope with alpha-actinin and a fast skeletal muscle fiber protein. Exp Cell Res 1996;225:306–14. 4. Ross R. Atherosclerosis—an inflammatory disease. N Engl J Med 1999;340:115–26. 5. Blindt R, Krott N, Hanrath P, et al. Expression patterns of integrins on quiescent and invasive smooth muscle cells and impact on cell locomotion. J Mol Cell Cardiol 2002;34:1633–44. 6. Nobes CD, Hall A. Rho GTPases control polarity, protrusion, and adhesion during cell movement. J Cell Biol 1999;144:1235–44. 7. Campbell GR, Campbell JH. The phenotypes of smooth muscle expressed in human atheroma. Ann N Y Acad Sci 1990;598:143–58. 8. Raines EW, Ross R. Smooth muscle cells and the pathogenesis of the lesions of atherosclerosis. Br Heart J 1993;69:S30–7. 9. Massberg S, Vogt F, Dickfeld T, et al. Activated platelets trigger an inflammatory response and enhance migration of aortic smooth muscle Figure 8 Regulation of SMC Migration by AAMP cells. Thromb Res 2003;110:187–94. 10. Yan Z, Hansson GK. Overexpression of inducible nitric oxide synthase Abbreviations as in Figure 2. Figure illustration by Rob Flewell. by neointimal smooth muscle cells. Circ Res 1998;82:21–9. 11. Blindt R, Bosserhoff AK, Dammers J, et al. Downregulation of N-cadherin in the neointima stimulates migration of smooth muscle decisive for the regulation of SMC migration. Thus, a block- cells by RhoA deactivation. Cardiovasc Res 2004;62:212–22. ade of membranous AAMP by an inhibitory antibody or by 12. Beckner ME, Peterson VA, Moul DE. Angio-associated migratory cell protein is expressed as an extracellular protein by blood-vessel- downregulation of AAMP decreases SMC migration and, associated mesenchymal cells. Microvasc Res 1999;57:347–52. concomitantly, results in a decreased activity of the small 13. Megens RT, Reitsma S, Schiffers PH, et al. Two-photon microscopy of GTPase RhoA, which is known to play a key role for control vital murine elastic and muscular arteries. Combined structural and of cellular migration (6,22). Computed protein structure anal- functional imaging with subcellular resolution. J Vasc Res 2007;44:87–98. 14. Blindt R, Bosserhoff AK, Zeiffer U, et al. Abciximab inhibits the ysis revealed a potential transmembrane domain between migration and invasion potential of human coronary artery smooth amino acids 322 and 345 and thereby supports the significance muscle cells. J Mol Cell Cardiol 2000;32:2195–206. of AAMP as a membrane-associated protein. Because homol- 15. Vogt F, Stein A, Rettemeier G, et al. Long-term assessment of a novel biodegradable paclitaxel-eluting coronary polylactide stent. Eur ogy analysis shows that AAMP shares sequence homology Heart J 2004;25:1330–40. with immunoglobulin superfamily members like cellular adhe- 16. Blindt R, Vogt F, Astafieva I, et al. A novel drug-eluting stent coated sion molecules NCAM, PECAM, and LFA-2, which are with an integrin-binding cyclic Arg-Gly-Asp peptide inhibits neoin- timal hyperplasia by recruiting endothelial progenitor cells. J Am Coll known to mediate adhesion and migration of various malig- Cardiol 2006;47:1786–95. nant circulating cells (1), a potential binding of AAMP to 17. Image J: Image Processing and Analysis in Java. Available at: http:// heparin or to beta-actinin was suggested. Applying advanced rsb.info.nih.gov/ij/. Accessed January 10, 2006. imaging and cell fractioning techniques, this study could clearly 18. Zernecke A, Schober A, Bot I, et al. SDF-1alpha/CXCR4 axis is instrumental in neointimal hyperplasia and recruitment of smooth demonstrate that AAMP is localized in the cellular membrane muscle progenitor cells. Circ Res 2005;96:784–91. and regulates SMC migration via the RhoA pathway. 19. Allander SV, Nupponen NN, Ringner M, et al. Gastrointestinal Figure 8 summarizes these findings schematically. Due to stromal tumors with KIT mutations exhibit a remarkably homoge- neous expression profile. Cancer Res 2001;61:8624–8. our data, AAMP can also be found in the cytosol but a 20. Adeyinka A, Emberley E, Niu Y, et al. Analysis of in functional role for the regulation of cell migration could not ductal carcinoma in situ of the breast. Clin Cancer Res 2002;8:3788–95. be attributed to cytosolic AAMP. Furthermore, secreted 21. Beckner ME, Jagannathan S, Peterson VA. Extracellular angio-associated migratory cell protein plays a positive role in angiogenesis and is regulated AAMP to the extracellular area as suggested previously by astrocytes in coculture. Microvasc Res 2002;63:259–69. could not be detected. 22. Raftopoulou M, Hall A. Cell migration: Rho GTPases lead the way. In summary, we have identified AAMP, abundantly Dev Biol 2004;265:23–32. produced by neointimal SMCs after vascular injury, as a y y y potential regulator of SMC migration in atherosclerotic and Key Words: atherosclerosis migration neointima formation restenosis y smooth muscle cells. restenotic disease. . Reprint requests and correspondence: Dr. Rüdiger Blindt, Depart- APPENDIX ment of Cardiology, RWTH Aachen University, Pauwelsstrasse 30, 52074 Aachen, Germany. E-mail: [email protected]. For supplementary material regarding the culture of smooth muscle cells, please see the online version of this article.