Multimodality Molecular Imaging Identifies Proteolytic and Osteogenic Activities in Early Aortic...

Multimodality Molecular Imaging Identifies Proteolytic andOsteogenic Activities in Early Aortic Valve Disease

Elena Aikawa, MD, PhD; Matthias Nahrendorf, MD; David Sosnovik, MD; Vincent M. Lok, BA;Farouc A. Jaffer, MD, PhD; Masanori Aikawa, MD, PhD; Ralph Weissleder, MD, PhD

Background—Visualizing early changes in valvular cell functions in vivo may predict the future risk and identifytherapeutic targets for prevention of aortic valve stenosis.

Methods and Results—To test the hypotheses that (1) aortic stenosis shares a similar pathogenesis to atherosclerosis and(2) molecular imaging can detect early changes in aortic valve disease, we used in vivo a panel of near-infraredfluorescence imaging agents to map endothelial cells, macrophages, proteolysis, and osteogenesis in aortic valves ofhypercholesterolemic apolipoprotein E–deficient mice (30 weeks old, n�30). Apolipoprotein E–deficient mice with noprobe injection (n�10) and wild-type mice (n�10) served as controls. Valves of apolipoprotein E–deficient micecontained macrophages, were thicker than wild-type mice (P�0.001), and showed early dysfunction detected by MRIin vivo. Fluorescence imaging detected uptake of macrophage-targeted magnetofluorescent nanoparticles (24 hours afterinjection) in apolipoprotein E–deficient valves, which was negligible in controls (P�0.01). Valvular macrophagesshowed proteolytic activity visualized by protease-activatable near-infrared fluorescence probes. Ex vivo magneticresonance imaging enhanced with vascular cell adhesion molecule-1–targeted nanoparticles detected endothelialactivation in valve commissures, the regions of highest mechanical stress. Osteogenic near-infrared fluorescence signalscolocalized with alkaline phosphatase activity and expression of osteopontin, osteocalcin, Runx2/Cbfa1, Osterix, andNotch1 despite no evidence of calcium deposits, which suggests ongoing active processes of osteogenesis in inflamedvalves. Notably, the aortic wall contained advanced calcification. Quantitative image analysis correlated near-infraredfluorescence signals with immunoreactive vascular cell adhesion molecule-1, macrophages, and cathepsin-B (P�0.001).

Conclusions—Molecular imaging can detect in vivo the key cellular events in early aortic valve disease, includingendothelial cell and macrophage activation, proteolytic activity, and osteogenesis. (Circulation. 2007;115:377-386.)

Key Words: valves � stenosis � inflammation � atherosclerosis � hypercholesterolemia � imaging

Aortic valve disease is a progressive disorder that rangesfrom mild valve thickening to severe calcification with

impaired leaflet motion or aortic valve stenosis.1 Aortic valvestenosis is the most common valvular heart disease; no effectivetherapy is currently available, however, other than surgical valvereplacement, performed in �85 000 patients in the United Statesand 275 000 worldwide annually.2,3 Clinical evidence suggeststhat coronary atherosclerosis and aortic valve stenosis sharesimilar epidemiological risk factors, such as age, sex, hypercho-lesterolemia, and hypertension.4–6 Clinicopathological studies ofhuman stenotic aortic valves identified lesions similar to those inatherosclerotic plaques that contained inflammatory cells andcalcific deposits.7,8 Preclinical studies further demonstrated ath-erosclerosis-like lesions in the aortic valve leaflets in rabbit andmouse models of atherosclerosis.9–11 The cellular and molecular

factors involved in the development of aortic valve stenosisremain largely obscure, however. Understanding the mecha-nisms of aortic valve inflammation and calcification will providemechanistic insights into the pathogenesis of aortic valve dis-ease. Furthermore, detection of early molecular and functionalabnormalities in aortic valves in vivo may predict the future riskof subclinical valvular lesions and identify targets for effectivetherapeutic strategies to prevent aortic valve stenosis.

Editorial p 297Clinical Perspective p 386

We previously demonstrated a common paradigm of inter-stitial and endothelial cell activation during valve develop-ment, disease, surgical transplantation, and tissue engineer-ing.12–17 Activated valvular cells overexpress matrix

Received August 3, 2006; accepted November 6, 2006.From the Center for Molecular Imaging Research (E.A., M.N., D.S., V.M.L., F.A.J., R.W.), Massachusetts General Hospital, Harvard Medical School,

Charlestown, Mass; Cardiology Division (D.S., F.A.J.), Department of Medicine, Massachusetts General Hospital, Boston, Mass; Cardiovascular Division(M.A.), Department of Medicine, Brigham and Women’s Hospital, Boston, Mass; and Donald W. Reynolds Cardiovascular Clinical Research Center(E.A., M.N., F.A.J., M.A., R.W.), Harvard Medical School, Boston, Mass.

Guest Editor for this article was Donald D. Heistad, MD.Correspondence to Elena Aikawa, MD, PhD, or Ralph Weissleder, MD, PhD, Center for Molecular Imaging Research, Massachusetts General Hospital,

Harvard Medical School, 149 13th St, Charlestown, MA 02129. E-mail [email protected] or [email protected]© 2007 American Heart Association, Inc.

Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.654913

377

Valvular Heart Disease

by guest on July 20, 2015http://circ.ahajournals.org/Downloaded from

http://circ.ahajournals.org/

metalloproteinases (eg, MMP-1, MMP-2, MMP-9, andMMP-13),12,13,16,17 cysteine proteases (cathepsin S and ca-thepsin K),12 adhesion molecules (vascular cell adhesionmolecule-1 [VCAM-1] and intercellular adhesion molecule-1),17 embryonic myosin,12,13,17 and interleukin-1�.12 Thesemolecules play various roles in inflammation and tissueremodeling and may represent surrogate end points for valvedevelopment, remodeling, and disease progression. Newerimaging technologies currently available may detect in vivoexpression and activity of molecules responsible for valvepathology and monitor maturation/remodeling of native andtissue-engineered valves in individual patients. We recentlyused such imaging approaches to demonstrate endothelial celland macrophage activation in aortas of hypercholesterolemicapolipoprotein E–deficient (apoE�/�) mice, an establishedmodel of atherosclerosis.18–20 We therefore hypothesized thatmultimodality molecular imaging can detect in vivo earlychanges in valvular cell functions and identify novel molec-ular targets for prevention of aortic valve stenosis. Thepresent study aimed to map key cellular and molecularfunctions, such as (1) endothelial cell activation, (2) macro-phage activation, (3) proteolytic enzyme activity, and (4)osteoblastic activity in the aortic valves of hypercholesterol-emic apoE�/� mice using a recently developed comprehensiveset of near-infrared fluorescent (NIRF) imaging probes thatspecifically target biological processes associated withinflammation.

MethodsAnimal ProtocolWe studied aortic valves of 30 apoE�/� mice at 30 weeks of age thathad received an atherogenic diet (Teklad TD 88137; 42% milk fat,0.2% total cholesterol, Harlan, Indianapolis, Ind) for 20 weeks.Age-matched wild-type C57/B6 mice (n�10, Jackson Laboratory,Bar Harbor, Me) and apoE�/� mice with no probe injection (n�10)served as controls. Serum cholesterol levels were determined at thetime of euthanasia with a colorimetric assay according to themanufacturer’s instructions (Raichem, San Diego, Calif). The Sub-committee on Research Animal Care at Massachusetts GeneralHospital approved all procedures.

Macroscopic Fluorescence Reflectance ImagingMice received imaging agents or saline (control) via intravenousinjection 24 or 48 hours before imaging. After mice had beeneuthanized, aortas were perfused with saline, dissected, and imagedto map the macroscopic NIRF signals elaborated from each imagingagent at the aortic root with a custom-built fluorescence reflectanceimaging system equipped with multichannel filter sets, includinggreen (green fluorescent protein/fluorescein isothiocyanate; excita-tion, 406 to 450 nm; emission, 495 to 525 nm), far red (VT680;excitation, 615 to 645 nm; emission, 680 to 720 nm), and near-infrared (indocyanine green; excitation, 716 to 756 nm; emission,780 to 820 nm; Omega Optical, Brattleboro, Vt). Fluorescenceimages were obtained with an exposure time of 1 to 60 seconds.Subsequently, aortic roots dissected through the area of high NIRFsignal were processed for histological analysis.

Microscopic Laser Scanning Fluorescence ImagingAfter the aortic root was isolated, multichannel fluorescence imagingwas performed with a laser scanning fluorescence microscope(IV100, Olympus Corp, Tokyo, Japan) specifically developed forimaging small experimental animals. Three laser lines at 450, 680,and 750 nm were used. Images of 512�512 pixels with a pixel sizeof �2.75�2.8 �m/pixel were collected with the FluoView 300

software program (Olympus) and stored as multilayer, 16-bit imagefile format (TIFF) files. Subsequently, aortic roots with leaflets wereprocessed for histological analyses.

Magnetic Resonance ImagingFor high-resolution ex vivo magnetic resonance imaging (MRI), theaortic root was immersed in liquid agar to minimize air–tissueinterface susceptibility artifacts and movement during imaging. Theaorta was then placed in a 14.0-T vertical-bore system with micro-imaging capabilities (Bruker BioSpin MRI, Inc, Billerica, Mass).Gradient-recalled echocardiographic images were obtained with thefollowing parameters: in-plane resolution, 70�70 mm; slice thick-ness, 0.5 mm; repetition time, 200 ms; and echo time. 7.0 ms. Imageswere analyzed with a shareware software package (OsiriX version1.7.1). After imaging, the aorta was removed from the agar andembedded for histological analyses.

In vivo MRI was performed under inhalation anesthesia (isoflu-rane 1% to 2% vol/vol plus 4 L of O2) on a 9.4-T horizontal-borescanner (Bruker Biospec, Billerica, Mass). We obtained bright-bloodcine images with ECG and respiratory gating (SA Instruments, StonyBrook, NY) using a gradient-echo FLASH sequence and a dedicatedmouse cardiac surface coil. Imaging parameters were as follows:in-plane resolution, 125�125 �m; slice thickness, 1 mm; 16 framesper RR interval (repetition time, 7.0 to 8.0 ms); echo time, 2.7 ms;and number of excitations, 8.

Correlative Histopathological Assessment

Morphological CharacterizationTissue samples were frozen in OCT compound (Sakura Finetech,Torrance, Calif), and 5-�m serial sections were cut through the aorticvalves. All 3 cusps were stained with hematoxylin and eosin forgeneral morphology. Activity of alkaline phosphatase (crucial forinitiating mineralization) was detected on cryosections that weredirectly incubated with conjugated antibody (red alkaline phospha-tase substrate kit; Vector Labs, Burlingame, Calif). Von Kossa silverstain was used to visualize inorganic phosphate calcium salts andAlizarin red to detect calcium deposits on adjacent sections. Aorticarches with prominent calcification were used as positive controls.

Multichannel Fluorescence MicroscopyWe performed multichannel fluorescence microscopy to examineactivation of valvular cells and proteolytic and osteogenic activity.Sections were imaged with an upright epifluorescence microscope(Eclipse 80i, Nikon Instruments, Melville, NY) with a cooled CCDcamera (Cascade, Photometrics, Tucson, Ariz). Fluorescence imageswere obtained at a wavelength of green (filter, 480�20 nm;excitation, 535�25 nm; emission, Q505LP bandpass), far red (filter,650�22.5 nm; excitation, 710�25 nm; emission, Q680LP band-pass), or near-infrared (filter, 775�25 nm; excitation, 845�27.5 nm;emission, Q810LP bandpass), depending on the probe. The sameexposure time, which ranged from 500 to 2000 ms, was used for eachprobe.

ImmunohistochemistryFor validation of NIRF signals, we used immunohistochemistry formacrophages (rat monoclonal antibody against mouse Mac3, BDBiosciences, San Jose, Calif), myofibroblasts (�-smooth muscleactin, 1A4, Dako, Carpinteria, Calif), activated endothelial cells(anti-mouse VCAM-1/CD106, BD Pharmingen, San Diego, Calif),gelatinases (polyclonal rabbit anti-mouse MMP-2 and MMP-9,Chemicon International, Temecula, Calif), cathepsin B and K (goatpolyclonal antibodies, Santa Cruz Biotechnology, Santa Cruz, Calif),osteoblast differentiation markers (goat polyclonal antibodies againstmouse osteocalcin and osteopontin, Abcam, Cambridge, Mass),osteogenic transcription factors and signaling (goat polyclonal anti-human Cbfa1/Runx2 antibody, RD Systems, Minneapolis, Minn;rabbit polyclonal antibody to Sp7/Osterix, Abcam, Cambridge,Mass; and rabbit polyclonal antibody to cleaved Notch1, CellSignaling Technology, Inc, Danvers, Mass). The avidin-biotin per-oxidase method was used for immunohistochemistry. The reaction

378 Circulation January 23, 2007



was visualized with a 3-amino-9-athyl-carbazol substrate (AEC,Sigma Chemical, St Louis, Mo), which yielded red reaction products.Adjacent sections treated with nonimmune IgG provided controls forantibody specificity. Images were captured with a digital camera(Nikon DXM 1200-F, Nikon Inc, Melville, NY).

Targeted and Activatable MolecularImaging Agents

Macrophage-Targeted Magnetofluorescent NanoparticlesMagnetofluorescent nanoparticles (MFNPs) are internalized by mac-rophages after systemic administration and therefore serve as imag-ing agents for in vivo detection of these proinflammatory phago-cytes.18 The nanoparticle was covered with a layer of dextran,contained a superparamagnetic iron oxide core detectable by MRI,and was labeled with far-red fluorochromes for fluorescence detec-tion (VT680; excitation/emission, 673/694 nm; Amersham Bio-sciences, Amersham, United Kingdom).

VCAM-1–Targeted Imaging AgentImaging of VCAM-1 expression in vivo used a peptide-derivatizednanoparticle (VCAM-1 internalizing nanoparticle-28) that triggerscellular internalization and “trapping” in VCAM-1–expressing cells,which leads to a biological signal amplification recently developedby in vivo phage display in apoE�/� mice.21,22 This linear peptide ishomologous to the integrin very late antigen-4, a known ligand ofVCAM-1. VCAM-1–targeted peptides were then conjugated withMFNPs with an average of 2 fluorochromes per nanoparticle(fluorescence in NIRF window excitation/emission 673/694 nm).

Protease-Activatable NIRF AgentsActivatable NIRF agents, well validated in a mouse model ofatherosclerosis,19,23 visualized activity of gelatinases (MMP-2/9;Gelsense680, VisEn Medical, Inc, Woburn, Mass) and cysteineproteases (predominantly cathepsin B; Prosense680, VisEn Medical,Inc) in inflamed aortic valves. These quenched substrate probesproduce negligible fluorescence at baseline because of closelyspaced fluorochromes. However, on protease-mediated cleavage andfluorochrome release, the NIRF signal increases by approximately200-fold. Both agents fluoresce in the NIRF window of excitation/emission 673/694 nm wavelength.

CalcificationRecent studies demonstrated that bisphosphonate-conjugated imag-ing agents can detect the functions of osteoblasts and calcificdeposition in bone and coronary arteries.24,25 OsteoSense750 (VisEnMedical, Inc) binds to sites of calcification in vivo, particularly tohydroxyapatite, and serves as an imaging agent for detection ofosteoblastic activity. This agent elaborates fluorescence detectablethrough the NIRF window (excitation/emission 750/780 nm).

Quantitative Assessment and Statistical AnalysisOverall thickness of the leaflets averaged over 5 equally distributedlength measurements throughout the valve for both wild-type andapoE�/� mice was quantified with imaging software (IPLab version3.9.3; Scanalytic, Inc, Rockville, Md). Digitized NIRF and immu-nohistochemistry images were analyzed with imaging software(IPLab version 3.9.3). NIRF images were segmented by thresholdinggrayscale intensity at a fixed level (20 000 arbitrary fluorescenceunits), and 5 regions of interest were randomly selected along eachleaflet. The same sections were then processed for immunohisto-chemistry (Mac3, VCAM-1, and cathepsin B), digitized, and seg-mented by thresholding red intensity at a fixed level (100 arbitraryunits). The percent-positive area per leaflet was calculated andcorrelated with average NIRF signal intensity of the same region ofinterest. The valve target-to-background ratio was calculated asvalve signal/adjacent aorta background signal.

Data are presented as mean�SEM. The Student (unpaired) t testwas performed with GraphPad Prism (version 4.00, GraphPadSoftware, San Diego, Calif). Probability values less than 0.05 wereconsidered significant.

The authors had full access to the data and take full responsibilityfor the integrity of the data. All authors have read and agree to themanuscript as written.

Results

Aortic Valves of Hypercholesterolemic Mice DevelopLesions Similar to Early Atherosclerotic PlaquesAortic valve leaflets of wild-type mice were thin and com-posed predominantly of quiescent fibroblast-like cells with nomacrophage infiltration (Figure 1A; serum cholesterol,108�26 mg/dL). In contrast, cholesterol-fed apoE�/� mice of30 weeks of age (serum cholesterol, 588�47 mg/dL;P�0.001 versus wild-type controls) had thickened leafletswith macrophage-rich subendothelial lesions on the aortic siteof the valve, located mostly in the base of the leaflet (Figure1A). Quantitative analysis revealed significantly thicker leaf-lets in apoE�/� mouse valves than in wild-type mice

Figure 1. Thickened valve leaflets in hypercholesterolemicapoE�/� mice. A, Aortic valve leaflets of wild-type mice (WT)were thin, with no macrophage infiltration. Leaflets of 30-week-old apoE�/� mice have an increased overall thickness associ-ated with the accumulation of the macrophages (Mac3) in thebase of the leaflet (arrow). Magnification �100. B, ApoE�/� mice(n�20) have a significant increase in overall leaflet thicknesscompared with wild type (n�10); P�0.0001. HE indicates hema-toxylin and eosin.

Aikawa et al Molecular Imaging of Aortic Valve Disease 379



(89.3�7.5 �m, n�20 versus 31.3�2.1 �m, n�10;P�0.0001; Figure 1B).

Endothelial Cell Activation Occurs in theCommissures of Diseased Aortic ValvesWe examined endothelial cell activation gauged by VCAM-1expression in aortic valves of apoE�/� mice that received theVCAM-1–targeted agent (VCAM-1 internalizingnanoparticle-28, 30 mg/kg iron, n�5) or 0.9% saline (n�3)via intravenous injection 48 hours before imaging. High-resolution 14.0-T MRI demonstrated signal enhancement exvivo. As visualized on short-axis slices through the aorticroot, VCAM-1 internalizing nanoparticle-28 predominantlytargeted the commissures of the aortic valve, providing arobust landmark for histopathological assessment (Figure2A). Fluorescence NIRF microscopy (680 nm) of VCAM-1internalizing nanoparticle-28 distribution visualized ex vivoby MRI correlated with VCAM-1 expression detected byimmunohistochemistry (R2�0.784, P�0.001; Figure 2B and2C).

Macrophages Accumulate in Early AorticValve LesionsA subset of apoE�/� mice received macrophage-targetedMFNPs (15 mg/kg iron, n�5) or 0.9% saline (n�3) 24 hours

before imaging. Fluorescence reflectance imaging throughthe 680-nm channel detected NIRF signals in the areascorresponding to the aortic root of mice that received MFNPs(Figure 3A). On cross sections through the aortic valve,immunoreactive macrophages colocalized with NIRF signals.In contrast, we found no significant NIRF signal enhance-ment in saline-injected controls either ex vivo or in situ(Figure 3B). The target-to-background ratio was 4-fold higherin apoE�/� mice than in controls (6.0 versus 1.6; P�0.01).Quantitative histopathological analysis correlated NIRF sig-nals (detected by fluorescence microscopy through the680-nm channel) with macrophages (detected by Mac3 stain-ing) in valvular lesions (R2�0.879, P�0.001; Figure 3C).

Macrophages and Activated ValvularMyofibroblasts Elaborate Proteolytic Activity inEarly Aortic Valve DiseaseWe further assessed the activity of matrix-degrading enzymesin aortic valves of cholesterol-fed apoE�/� mice injected withNIRF protease-activatable probes: cysteine proteases (cathep-sin B, n�5) or gelatinases (MMP-2 and MMP-9, n�5).Wild-type control mice produced undetectable levels of NIRFsignals determined by macroscopic fluorescence reflectanceimaging (data not shown) and multichannel fluorescence

Figure 2. Endothelial cell activationoccurs in the commissures of diseasedaortic valves. A, Ex vivo MRI. Left, Long-axis view demonstrated the aortic archand root. Dotted line demonstrates sliceposition of short-axis view. Middle,Short-axis view shows negative signalenhancement (darkening) caused byuptake of VCAM-1–targeted nanopar-ticles. Right, Color-coded signal intensi-ties (red) show focused uptake ofVCAM-1 in commissures (arrows). B,Immunoreactive VCAM-1 colocalizeswith NIRF signal (excitation/emission673/694 nm, exposure time 500 ms) inaortic valve commissure. Magnification�400. C, Quantitative image analysiscorrelates immunoreactive VCAM-1 withNIRF signals.




microscopy (Figure 4A). Notably, normal valves in wild-typemice did not contain immunoreactive cathepsin B. Fluores-cence reflectance imaging in apoE�/� mice injected with thecathepsin-activatable NIRF probe yielded strong signals atthe level of aortic valves (Figure 4B). Correlative histopatho-logical analysis colocalized NIRF signals with immunoreac-tive cathepsin B in the macrophage-rich lesions (Mac3) ofaortic valves (Figure 4C). Quantitative analysis further dem-onstrated a close positive correlation between NIRF-positiveareas and cathepsin B–immunopositive areas per leaflet(R2�0.858, P�0.001; Figure 4D). Sites of atheroscleroticvalves of apoE�/� mice that exhibited fluorescence signalselaborated from an activatable NIRF probe for gelatinasescolocalized with activated myofibroblast-like cells (�-smooth

muscle actin–positive cells) and macrophages (Mac3) bearingimmunoreactive MMP-2 and MMP-9 (Figure 5).

Valvular Myofibroblasts Exhibit OsteoblasticActivity in the Early Stage of Aortic Valve StenosisTo detect osteoblastic activity in inflamed aortic valves,we administered OsteoSense750 in apoE�/� mice 24 hoursbefore imaging (n�10). Mice were coinjected with spec-trally distinct NIRF MFNPs to visualize macrophages. Exvivo imaging with multichannel, high-resolution, laserscanning fluorescence microscopy on opened aortic rootdetected signals for osteogenic activity elaborated fromOsteoSense750 (750 nm, target-to-background ra-tio�29.7�1.6, P�0.01; Figure 6A). A macrophage-

Figure 3. Macrophages accumulate inearly aortic valve lesions of apoE�/�

mice. A, Macroscopic fluorescent reflec-tance imaging exhibited NIRF signal inthe aortic root of mice that receivedMFNPs (top left; VT680; excitation, 615to 645 nm; emission, 680 to 720 nm).Histopathological analysis with mul-tichannel fluorescence microscopy (exci-tation/emission, 673/694 nm, MFNP; 480nm, autofluorescence; exposure time,500 ms), hematoxylin and eosin (HE),and immunohistochemistry correlatedstrong NIRF MFNP (red fluorescence onmerged image) signal with the macro-phage-rich valvular lesions determinedby Mac3. Negligible autofluorescencewas detected through the 480-nm chan-nel (green fluorescence on mergedimage). Magnification �200. B, ApoE�/�

valves without probe injection showedno significant NIRF signal enhancementdetected by macroscopic (left) andmicroscopic (right) fluorescence imaging.Magnification �200. C, Quantitativeanalysis demonstrated a positive correla-tion between areas of NIRF and Mac3staining.




targeted NIRF MFNP (680 nm, target-to-background ra-tio�5.1�0.6, P�0.01) visualized inflammation (Figure6A). Fluorescence microscopy on cross sections of thesame aortic valve further colocalized NIRF signals forosteoblastic activity (750 nm) in myofibroblasts withalkaline phosphatase activity, immunoreactive osteopontinand osteocalcin (differentiated osteoblasts), Runx2/Cbfa1and Osterix (osteogenic transcription factors), and cleavedNotch1 (indicative of activated Notch signaling), whereas

Alizarin red stain and von Kossa stain detected no evi-dence of mineralization (Figure 6B), which suggests theactive ongoing processes of osteogenesis in inflamedvalves (detected by Mac3; data not shown). In contrast,cross sections through aortic arches of the same animalsdemonstrated prominent calcification detected by bothAlizarin red and von Kossa methods. Runx2/Cbfa1, Os-terix, and cleaved Notch1 were undetectable in aorticvalves of wild-type mice (data not shown).

Figure 4. Macrophages contribute toelastolytic activity in early aortic valvedisease. A, Cross sections of wild-type(WT) control valves elaborating negligibleNIRF signal enhancement were notimmunoreactive to cathepsin B. B, Mac-roscopic fluorescence imaging exhibitedNIRF signals in the aortic root of apoE�/�

mice that received cathepsin-activatableagent (VT680; excitation, 615 to 645 nm;emission, 680 to 720 nm). C, Correlativehistopathological analysis with mul-tichannel NIRF microscopy (excitation/emission, 673/694 nm; exposure time,1000 ms) and immunohistochemistrycolocalized NIRF signals with macro-phages (Mac3) expressing cathepsin B.Bottom panels demonstrate high-powerviews. Magnification �100 (top) and�400 (bottom). D, Quantitative analysiscorrelated NIRF signals and immunore-active cathepsin B.




MRI Detected In Vivo Early Valvular Dysfunctionin Aortic Valve DiseaseHigh-resolution cine magnetic resonance images examined invivo the functional integrity of the aortic valve. Threeconsecutive early diastolic frames from one such cine dem-onstrated that a jetlike area of flow-related spin dephasing,which produces a loss of signal, propagates across the leafletsof the aortic valve into the left ventricle (Figure 7). Thesefindings are consistent with aortic valve regurgitation, previ-ously described in apoE�/� mice by echocardiography.10,11

DiscussionAccumulating preclinical and clinical evidence suggests thataortic valve stenosis and atherosclerosis share similar mech-anisms, genetic mutations, risk factors, and histopathologicalfeatures (eg, macrophage accumulation and calcification).7–9

A recent study by Tanaka et al10 reported that aged apoE�/�

mice (95 weeks old) on regular chow developed degenerativevalve sclerosis that resembles human aortic stenosis. Thework by Drolet and colleagues11 demonstrated significantaortic valve abnormalities in low-density lipoprotein recep-tor–deficient mice fed a high-fat/high-carbohydrate diet. Thepresent study has further explored the pathogenesis of aorticvalve stenosis and the feasibility of imaging early changes inaortic valve disease. The present study demonstrated that a

high-cholesterol diet induced valve thickening in 30-week-old apoE�/� mice, which was associated with the presence ofthe macrophage-rich lesions that are highly prone to calcifi-cation (Figures 1 through 6); therefore, these mice likely havean accelerated rate of disease progression and can serve as amuch-needed animal model to study the pathogenesis ofaortic valve stenosis. The present study also visualized keycellular and molecular events (eg, endothelial cell activation,macrophage activation, proteolytic activity, and osteoblasticactivity/premineralization) in early lesions that may be in-volved in the development of valvular disease and causestructural changes resulting in valve insufficiency (Figure 7).Moreover, it unraveled dynamic biological processes in vivoin valve disease progression rather than merely analyzing thestatic and degenerative conditions at the time of death. Inaddition, the present study linked hypercholesterolemia andvalvular inflammation, which suggests that atherogenic fac-tors also contribute to the pathogenesis of aortic valvesclerosis.

Aortic valves open and close �100 000 times a day andtherefore bear repetitive changes in shape and dimensionthroughout the cardiac cycle. The flexion area of the aorticleaflets near the attachment of the aortic root (commissure)encounters the highest mechanical forces,26 which mightinduce endothelial cell activation/injury and expression ofadhesion molecules. Indeed, several studies demonstratedincreased VCAM-1, intercellular adhesion molecule-1, andE-selectin expression in surgically removed diseased heartvalves.27,28 Strict temporal and spatial regulation makes theseadhesion molecules ideal targets for molecular imaging. Thepresent study using MRI and NIRF microscopy clearlydemonstrates ex vivo that distribution of a VCAM-1–targetedagent mostly occurs in valve commissures and correlates wellwith immunoreactive VCAM-1 (Figure 2). These resultsfurther suggest that endothelial cell activation/damage occursat the regions of high flexure and increased mechanical forcesand that inflammatory cells likely enter the leaflets viacirculation in response to endothelial cell activation or injury.

In addition to mechanical stresses, various atherogeniccomponents such as elevated plasma lipids may induce valveendothelial cell activation that results in an amplificationcascade of events such as monocyte recruitment, visualized inthe present study by macrophage-targeted MFNPs. In addi-tion, we and others have shown that macrophages andvalvular interstitial cells (activated myofibroblast-like cells)elaborate excessive levels of proteolytic enzymes (eg, colla-genase-1/MMP-1, collagenase-3/MMP-13, gelatinase-A/MMP-2, and gelatinase-B/MMP-9) and cysteine endopro-teases (cathepsins) and contribute critically to collagen andelastin degradation that leads to vascular and valvular remod-eling and subsequent structural changes.12,13,15,16,29–31 Fur-thermore, elastolytic enzymes (eg, gelatinases, stromelysin,and cathepsins) may initiate elastin degradation, which pro-vides a nidus for hydroxyapatite crystal formation,32 andtherefore, these enzymes may play a considerable role inaortic calcification. Several lines of evidence also suggestosteopontin-induced MMP-9 activity in aortic mesenchymalcells, which may contribute to the vascular inflammation andcalcification.33 In the present study, we demonstrated os-

Figure 5. Macrophages and activated valvular myofibroblast-likecells elaborate gelatinolytic activity in early aortic valve disease.NIRF signal (excitation/emission 673/694 nm, red fluorescenceon merged image; 480 nm, green autofluorescence on mergedimage; exposure time 2000 ms; magnification �100) in aorticvalves of apoE�/� mice injected with gelatinase MMP-2/9 agentcolocalizes with MMP-2 and MMP-9 expressed by activatedmyofibroblasts (�-smooth muscle actin–positive cells) and mac-rophages (Mac3) in the valve leaflets. Magnification �400.




teopontin expression and MMP-9 activity in inflamed aorticvalves. In addition to extracellular matrix, MMPs cleavevarious substrates (eg, interleukin-1� precursor and tissuefactor pathway inhibitor) and therefore may enhance valvularinflammation.34,35 The multimodality imaging approach usedin the present study visualized proteolytic activity elaboratedfrom valvular macrophages and activated myofibroblasts

(Figures 3 through 5) and thus may provide a biologicalreadout of inflammation and matrix degradation and predictthe risk of subclinical aortic valve stenosis.

The high morbidity and mortality rates in patients withaortic sclerosis and the significant portion of those whosubsequently develop aortic stenosis suggest the need forclose follow-up and serial evaluation once lesions are iden-

Figure 6. Valvular myofibroblasts exhibitosteoblastic activity in the early stage ofaortic valve stenosis. A, Ex vivo imagingwith multichannel laser scanning fluores-cence microscopy performed on theexcised aortic root as shown on theschematic diagram. Multicolor fluores-cence imaging simultaneously visualized2 different biological processes: osteo-genesis (excitation/emission, 750 nm,red) and inflammation (excitation/emis-sion, 680 nm, green). B, Correlative his-topathological analysis with multichannelNIRF microscopy detected osteoblasticactivity (excitation/emission, 750/780 nm;red fluorescence on merged image;exposure time, 500 ms) and inflammation(excitation/emission, 673/694 nm; greenfluorescence on merged image; expo-sure time, 500 ms) on cross sectionsthrough the leaflets. NIRF osteogenicsignals colocalized with the alkalinephosphatase activity (ALP) elaborated byactivated myofibroblast-like cells(�-smooth muscle actin), osteocalcin andosteopontin, osteogenic transcriptionfactors Runx2/Cbfa1 and Osterix, andcleaved Notch1, which suggests theactive processes of ongoing mineraliza-tion. Notably, von Kossa and Alizarin redstain for calcium deposition showed noevidence of microscopic mineralization inthe leaflets, whereas cross sectionsthrough the aortic arch showed promi-nent calcification. Magnification �400.

Figure 7. MRI detected in vivo early val-vular dysfunction in aortic valve disease.MR images of 3 successive frames froma series of 16 frames in early diastoleacquired at long-axis view of the leftventricular outflow tract and the aortademonstrate aortic regurgitation. Theregurgitation jet originates at the valveleaflets and reaches the midventricularlevel. The jet causes flow-related spindephasing, which leads to the typicalblack signal (arrows).




tified.36 Conventional structural imaging modalities can iden-tify prominent late-stage calcification; no current imagingmethods can detect in vivo early mineralization and osteo-genesis in cardiac valves, however. The present study de-tected an amplified osteoblastic NIRF signal that resultedfrom binding of the imaging agent to nanomolar concentra-tions of calcium/hydroxyapatite complexes elaborated by�-smooth muscle actin–positive cells.37 These valvularmyofibroblast-like cells exhibited alkaline phosphatase activ-ity, osteopontin, osteocalcin, and osteogenic transcriptionfactors such as Runx2/Cbfa1 and Osterix, which indicateshighly regulated active processes of ongoing mineralizationbefore development of macroscopic and microscopic calcifi-cation (Figure 6). We also found that large populations ofmyofibroblast-like cells contained a cleaved form of Notch1,which may direct osteoblast differentiation.38 In addition, aspectrally distinct NIRF signal in the same valves colocalizedwith macrophages that expressed osteoclastic cathepsin K(expressed during bone resorption; data not shown),39 whichsuggests that the early development of aortic valve stenosisinvolves osteoblastic and osteoclastic activities in parallel.What causes the imbalance of these activities toward miner-alization in early aortic valve disease, however, requiresfurther investigation. Nevertheless, the results of the presentstudy documented here show an association of valvularlesions with features of typical of atherosclerotic plaques,including endothelial activation, inflammation, proteolyticactivity, and osteogenesis. Therefore, modification of athero-genic factors and pharmacological therapies that target proin-flammatory pathways may retard the progression of aorticvalve calcification when introduced early. Molecular imagingof the earliest stages of calcification may identify high-riskvalves while disease is silent and may enable the monitoringof valvular osteogenic activity during therapeutic interven-tions such as lipid lowering.

Collectively, our present findings visualizing early changesof cellular functions in vivo support a concept of thedevelopment of calcific aortic valve disease (see review byO’Brien38). Mechanical forces and oxidized lipids may acti-vate valvular endothelium and initiate recruitment of inflam-matory monocytes/macrophages. Activated macrophagesproduce a variety of cytokines, growth factors, and matrix-degrading enzymes. Proteolytic enzyme action, in turn, maycause extracellular matrix remodeling and thickening/stiffen-ing of the leaflets, resulting in valvular dysfunction. Theresultant altered mechanical stresses and disturbed flowpatterns may further induce inflammation and differentiationof valvular fibroblasts into activated myofibroblasts and,subsequently, into osteoblast-like cells through upregulationof the Runx2/Cbfa1 pathway, leading to the deposition ofcalcium primarily in the valvular commissures and to immo-bilization of the aortic leaflets. Therefore, a better under-standing of the molecular mechanisms of calcification thatcauses dysfunction of aortic valves is required to improvetherapies for aortic valve stenosis.40 Ongoing efforts thatcombine rigorous assessment of valve pathobiology andfurther development of imaging technologies will providenovel insights into an optimum therapeutic strategy forpatients with aortic valve stenosis and other valvular diseases.

AcknowledgmentsWe thank colleagues at the Center for Molecular Imaging Research,Massachusetts General Hospital, including Drs Kimberly Kelly,Ching-Hsuan Tung, and Nikolay Sergeyev for the developmentand/or synthesis of imaging agents, Jose-Luiz Figueiredo for hisexcellent surgical skills, and Rainer Kohler for assisting withlaser-scanning fluorescent microscopy. We thank Guangping Daiand Christian Farrar, Martinos Center for Biomedical Imaging, fortheir technical support. We also acknowledge Dr Peter Libby,Brigham and Women’s Hospital, for helpful discussions.

Sources of FundingThe present study was supported in part by grants from the NationalHeart, Lung, and Blood Institute (UOI-HL080731 and ROI-HL078641 to Dr Weissleder) and the Donald W. Reynolds Founda-tion (Dr Weissleder).

DisclosuresNone.

References1. Freeman RV, Otto CM. Spectrum of calcific aortic valve disease: patho-

genesis, disease progression, and treatment strategies. Circulation. 2005;111:3316–3326.

2. Rabkin-Aikawa E, Mayer JE Jr, Schoen FJ. Heart valve regeneration. AdvBiochem Eng Biotechnol. 2005;94:141–179.

3. Badylak SF. Regenerative medicine approach to heart valve replacement.Circulation. 2005;111:2715–2716.

4. Stewart BF, Siscovick D, Lind BK, Gardin JM, Gottdiener JS, Smith VE,Kitzman DW, Otto CM. Clinical factors associated with calcific aorticvalve disease: Cardiovascular Health Study. J Am Coll Cardiol. 1997;29:630–634.

5. Allison MA, Cheung P, Criqui MH, Langer RD, Wright CM. Mitral andaortic annular calcification are highly associated with systemic calcifiedatherosclerosis. Circulation. 2006;113:861–866.

6. Mohler ER III. Mechanisms of aortic valve calcification. Am J Cardiol.2004;94:1396–1402.

7. Otto CM, Kuusisto J, Reichenbach DD, Gown AM, O’Brien KD. Char-acterization of the early lesion of “degenerative” valvular aortic stenosis:histological and immunohistochemical studies. Circulation. 1994;90:844–853.

8. Olsson M, Thyberg J, Nilsson J. Presence of oxidized low densitylipoprotein in nonrheumatic stenotic aortic valves. Arterioscler ThrombVasc Biol. 1999;19:1218–1222.

9. Rajamannan NM, Subramaniam M, Springett M, Sebo TC, Niekrasz M,McConnell JP, Singh RJ, Stone NJ, Bonow RO, Spelsberg TC. Atorva-statin inhibits hypercholesterolemia-induced cellular proliferation andbone matrix production in the rabbit aortic valve. Circulation. 2002;105:2660–2665.

10. Tanaka K, Sata M, Fukuda D, Suematsu Y, Motomura N, Takamoto S,Hirata Y, Nagai R. Age-associated aortic stenosis in apolipoproteinE–deficient mice. J Am Coll Cardiol. 2005;46:134–141.

11. Drolet MC, Roussel E, Deshaies Y, Couet J, Arsenault M. A high fat/highcarbohydrate diet induces aortic valve disease in C57BL/6J mice. J AmColl Cardiol. 2006;47:850–855.

12. Rabkin E, Aikawa M, Stone JR, Fukumoto Y, Libby P, Schoen FJ.Activated interstitial myofibroblasts express catabolic enzymes andmediate matrix remodeling in myxomatous heart valves. Circulation.2001;104:2525–2532.

13. Rabkin E, Hoerstrup SP, Aikawa M, Mayer JE Jr, Schoen FJ. Evolutionof cell phenotype and extracellular matrix in tissue-engineered heartvalves during in-vitro maturation and in-vivo remodeling. J Heart ValveDis. 2002;11:308–314.

14. Rabkin E, Schoen FJ. Cardiovascular tissue engineering. CardiovascPathol. 2002;11:305–317.

15. Rabkin-Aikawa E, Farber M, Aikawa M, Schoen FJ. Dynamic andreversible changes of interstitial cell phenotype during remodeling ofcardiac valves. J Heart Valve Dis. 2004;13:841–847.

16. Rabkin-Aikawa E, Aikawa M, Farber M, Kratz JR, Garcia-Cardena G,Kouchoukos NT, Mitchell MB, Jonas RA, Schoen FJ. Clinical pulmonaryautograft valves: pathologic evidence of adaptive remodeling in the aorticsite. J Thorac Cardiovasc Surg. 2004;128:552–561.




17. Aikawa E, Whittaker P, Farber M, Mendelson K, Padera RF, Aikawa M,Schoen FJ. Human semilunar cardiac valve remodeling by activated cellsfrom fetus to adult: implications for postnatal adaptation, pathology, andtissue engineering. Circulation. 2006;113:1344–1352.

18. Jaffer FA, Nahrendorf M, Sosnovik D, Kelly KA, Aikawa E, WeisslederR. Cellular imaging of inflammation in atherosclerosis using magnet-ofluorescent nanomaterials. Mol Imaging. 2006;5:85–92.

19. Deguchi JO, Aikawa M, Tung CH, Aikawa E, Kim DE, Ntziachristos V,Weissleder R, Libby P. Inflammation in atherosclerosis: visualizingmatrix metalloproteinase action in macrophages in vivo. Circulation.2006;114:55–62.

20. Pande AN, Kohler RH, Aikawa E, Weissleder R, Jaffer FA. Detection ofmacrophage activity in atherosclerosis in vivo using multichannel, high-resolution laser scanning fluorescence microscopy. J Biomed Opt. 2006;11:21009–21015.

21. Nahrendorf M, Jaffer FA, Kelly KA, Sosnovik D, Aikawa E, Libby P,Weissleder R. Noninvasive vascular cell adhesion molecule-1 imagingidentifies activated endothelium in atherosclerosis. Circulation. 2006;114:1504–1511.

22. Kelly KA, Allport JR, Tsourkas A, Shinde-Patil VR, Josephson L,Weissleder R. Detection of vascular adhesion molecule-1 expressionusing a novel multimodal nanoparticle. Circ Res. 2005;96:327–336.

23. Chen J, Tung CH, Mahmood U, Ntziachristos V, Gyurko R, FishmanMC, Huang PL, Weissleder R. In vivo imaging of proteolytic activity inatherosclerosis. Circulation. 2002;105:2766–2771.

24. Zaheer A, Lenkinski RE, Mahmood A, Jones AG, Cantley LC, FrangioniJV. In vivo near-infrared fluorescence imaging of osteoblastic activity.Nat Biotechnol. 2001;19:1148–1154.

25. Zaheer A, Murshed M, De Grand AM, Morgan TG, Karsenty G,Frangioni JV. Optical Imaging of hydroxyapatite in the calcified vascu-lature of transgenic animals. Arterioscler Thromb Vasc Biol. 2006;26:1132–1136.

26. Davies MJ, Treasure T, Parker DJ. Demographic characteristics ofpatients undergoing aortic valve replacement for stenosis: relation tovalve morphology. Heart. 1996;75:174–178.

27. Muller AM, Cronen C, Kupferwasser LI, Oelert H, Muller KM,Kirkpatrick CJ. Expression of endothelial cell adhesion molecules onheart valves: up-regulation in degeneration as well as acute endocarditis.J Pathol. 2000;191:54–60.

28. Ghaisas NK, Foley JB, O’Briain DS, Crean P, Kelleher D, Walsh M.Adhesion molecules in nonrheumatic aortic valve disease: endothelialexpression, serum levels and effects of valve replacement. J Am CollCardiol. 2000;36:2257–2262.

29. Aikawa M, Rabkin E, Okada Y, Voglic SJ, Clinton SK, Brinckerhoff CE,Sukhova GK, Libby P. Lipid lowering by diet reduces matrix metallo-

proteinase activity and increases collagen content of rabbit atheroma: apotential mechanism of lesion stabilization. Circulation. 1998;97:2433–2444.

30. Deguchi JO, Aikawa E, Libby P, Vachon JR, Inada M, Krane SM,Whittaker P, Aikawa M. Matrix metalloproteinase-13/collagenase-3deletion promotes collagen accumulation and organization in mouse ath-erosclerotic plaques. Circulation. 2005;112:2708–2715.

31. Helske S, Syvaranta S, Lindstedt KA, Lappalainen J, Oorni K, MayranpaaMI, Lommi J, Turto H, Werkkala K, Kupari M, Kovanen PT. Increasedexpression of elastolytic cathepsins S, K, and V and their inhibitorcystatin C in stenotic aortic valves. Arterioscler Thromb Vasc Biol.2006;8:1791–1798.

32. Qin X, Corriere MA, Matrisian LM, Guzman RJ. Matrix metallopro-teinase inhibition attenuates aortic calcification. Arterioscler ThrombVasc Biol. 2006;26:1510–1516.

33. Lai CF, Seshadri V, Huang K, Shao JS, Cai J, Vattikuti R, SchumacherA, Loewy AP, Denhardt DT, Rittling SR, Towler DA. Anosteopontin-NADPH oxidase signaling cascade promotes pro-matrix met-alloproteinase 9 activation in aortic mesenchymal cells. Circ Res. 2006;98:1479–1489.

34. Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors ofmetalloproteinases: structure, function, and biochemistry. Circ Res. 2003;92:827–839.

35. Belaaouaj AA, Li A, Wun TC, Welgus HG, Shapiro SD. Matrix metal-loproteinases cleave tissue factor pathway inhibitor: effects on coagu-lation. J Biol Chem. 2000;275:27123–27128.

36. Cosmi JE, Kort S, Tunick PA, Rosenzweig BP, Freedberg RS, Katz ES,Applebaum RM, Kronzon I. The risk of the development of aorticstenosis in patients with “benign” aortic valve thickening. Arch InternMed. 2002;162:2345–2347.

37. Reynolds JL, Joannides AJ, Skepper JN, McNair R, Schurgers LJ,Proudfoot D, Jahnen-Dechent W, Weissberg PL, Shanahan CM. Humanvascular smooth muscle cells undergo vesicle-mediated calcification inresponse to changes in extracellular calcium and phosphate concen-trations: a potential mechanism for accelerated vascular calcification inESRD. J Am Soc Nephrol. 2004;15:2857–2867.

38. O’Brien KD. Pathogenesis of calcific aortic valve disease: a diseaseprocess comes of age (and a good deal more). Arterioscler Thromb VascBiol. 2006;26:1721–1728.

39. Bossard MJ, Tomaszek TA, Thompson SK, Amegadzie BY, Hanning CR,Jones C, Kurdyla JT, McNulty DE, Drake FH, Gowen M, Levy MA.Proteolytic activity of human osteoclast cathepsin K: expression, purifi-cation, activation, and substrate identification. J Biol Chem. 1996;271:12517–12524.

40. Demer LL. Cholesterol in vascular and valvular calcification. Circulation.2001;104:1881–1883.

CLINICAL PERSPECTIVEAortic valves with severe stenosis and calcification that may cause clinical complications often are surgically replaced.Although current valve substitutes behave similarly in most recipients, responses of engineered tissue implantation mayhave considerable patient-to-patient variability. Therefore, it is important not only to understand how individual factors andresponses to injury affect the structure and function of tissue-engineered valves, but also to monitor, noninvasively, tissuehealing and remodeling in vivo, which requires emerging technologies such as molecular imaging, which is able to assesstissue composition, microstructure, and biological processes. The agents used in the present study can potentially be testedin a clinical setting and detected by high-resolution magnetic resonance imaging or optical approaches. Catheter-basedfluorescence sensors and confocal microscopy are expected to allow high-resolution, multichannel, in vivo imaging of thedifferent biological processes such as inflammation and osteogenesis. In addition, imaging of inflammatory and calcificfoci in the earliest aortic valve lesions will permit testing of pathophysiological hypotheses in vivo and could serve as anintermediate end point based on biological function in addition to anatomy in the evaluation of novel therapies that targetinflammation and calcification and, hence, permit intervention before aortic stenosis develops. Our findings suggest thatmolecular imaging approaches will (1) improve our understanding of the molecular bases of development of aortic stenosisand other valve diseases, (2) identify subclinical valvular lesions, (3) predict the future risk of such lesions, (4) help toestablish individualized therapeutic strategies, (5) evaluate effects of novel therapies, and (6) monitor remodeling ofimplanted tissue-engineered heart valves.




Masanori Aikawa and Ralph WeisslederElena Aikawa, Matthias Nahrendorf, David Sosnovik, Vincent M. Lok, Farouc A. Jaffer,

Aortic Valve DiseaseMultimodality Molecular Imaging Identifies Proteolytic and Osteogenic Activities in Early

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2007 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/CIRCULATIONAHA.106.654913

2007;115:377-386; originally published online January 15, 2007;Circulation.

http://circ.ahajournals.org/content/115/3/377World Wide Web at:

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

http://circ.ahajournals.org//subscriptions/

is online at: Circulation 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 this process is available in the

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

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


http://circ.ahajournals.org/content/115/3/377

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/


Multimodality Molecular Imaging Identifies Proteolytic and Osteogenic Activities in Early Aortic...

Documents

Transcript of Multimodality Molecular Imaging Identifies Proteolytic and Osteogenic Activities in Early Aortic...