Gene expression levels of matrix metalloproteinases in human atherosclerotic plaques and evaluation of radiolabeled inhibitors as imaging agents for plaque vulnerability

Nuclear Medicine and Biology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Nuclear Medicine and Biology

j ourna l homepage: www.e lsev ie r .com/ locate /nucmedbio

Gene expression levels of matrix metalloproteinases in humanatherosclerotic plaques and evaluation of radiolabeled inhibitors asimaging agents for plaque vulnerability

Adrienne Müller a,⁎, Stefanie D. Krämer a, Romana Meletta a, Katharina Beck a, Svetlana V. Selivanova a,Zoran Rancic d, Philipp A. Kaufmann b, Bernhard Vos c, Jörg Meding c, Timo Stellfeld c, Tobias K. Heinrich c,Marcus Bauser c, Joachim Hütter c, Ludger M. Dinkelborg c, Roger Schibli a, Simon M. Ametamey a

a Center for Radiopharmaceutical Sciences, ETH Zurich, Zurich, Switzerlandb Department of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerlandc Global Drug Discovery, Bayer Healthcare, Berlin, Germanyd Clinic for Cardiovascular Surgery, University Hospital Zurich, Zurich, Switzerland

a b s t r a c ta r t i c l e i n f o

⁎ Corresponding author at: Center for RadiopharmaceUSZ, Department of Chemistry and Applied BiosciencesPrelog-Weg 3, CH-8093 Zurich, Switzerland. Tel.: +4633 13 67.

E-mail address: [email protected] (A

http://dx.doi.org/10.1016/j.nucmedbio.2014.04.0850969-8051/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Müller A., et al, Gof radiolabeled inhibitors as imaging agents

Article history:
Received 28 February 2014Received in revised form 10 April 2014Accepted 12 April 2014Available online xxxx
Keywords:AtherosclerosisMatrix metalloproteinasesTissue inhibitor of metalloproteinase-3Carotid artery plaqueInflammationAutoradiography

Introduction: Atherosclerotic plaque rupture is the primary cause for myocardial infarction and stroke. Duringplaque progression macrophages and mast cells secrete matrix-degrading proteolytic enzymes, such asmatrix metalloproteinases (MMPs). We studied levels of MMPs and tissue inhibitor of metalloproteinases-3(TIMP-3) in relation to the characteristics of carotid plaques. We evaluated in vitro two radiolabeled probestargeting active MMPs towards non-invasive imaging of rupture-prone plaques.Methods: Human carotid plaques obtained from endarterectomy were classified into stable and vulnerable byvisual and histological analysis. MMP-1, MMP-2, MMP-8, MMP-9, MMP-10, MMP-12, MMP-14, TIMP-3, andCD68 levels were investigated by quantitative polymerase chain reaction. Immunohistochemistry was used tolocalize MMP-2 and MMP-9 with respect to CD68-expressing macrophages. Western blotting was applied todetect their active forms. A fluorine-18-labeled MMP-2/MMP-9 inhibitor and a tritiated selective MMP-9inhibitor were evaluated by in vitro autoradiography as potential lead structures for non-invasive imaging.Results: Gene expression levels of all MMPs and CD68 were elevated in plaques. MMP-1, MMP-9, MMP-12 and
MMP-14 were significantly higher in vulnerable than stable plaques. TIMP-3 expression was highest in stableand low in vulnerable plaques. Immunohistochemistry revealed intensive staining of MMP-9 in vulnerableplaques. Western blotting confirmed presence of the active form in plaque lysates. In vitro autoradiographyshowed binding of both inhibitors to stable and vulnerable plaques.Conclusions: MMPs differed in their expression patterns among plaque phenotypes, providing possibleimaging targets. The two tested MMP-2/MMP-9 andMMP-9 inhibitors may be useful to detect atheroscleroticplaques, but not the vulnerable lesions selectively.
© 2014 Elsevier Inc. All rights reserved.

1. Introduction

Ischemic stroke represents a major health problem and is animportant cause of long-term disability in several developed countries[1]. Mortality rate from stroke ranges from 10% to 30%, and itssurvivors remain at a high risk of recurrent ischemic events andmortality, both from myocardial infarction and repeated stroke [2].Rupture of atherosclerotic plaques and consequent thrombus forma-tion are considered as the crucial steps in the development of acute

utical Sciences of ETH, PSI and, ETH Hönggerberg, Vladimir-1 44 633 60 84; fax: +41 44

. Müller).

ene expression levels of matrifor plaque vulnerability, Nucl

cardiovascular events, such as cerebral and myocardial infarctions[3,4]. High-grade internal carotid artery stenosis is, therefore, almostalways an indication for immediate carotid endarterectomy (CEA) toavoid rupture of a potentially vulnerable plaque. Arterial stenosis isdiagnosed by duplex ultrasoundmagnetic resonance angiography andcomputerized tomographic angiography. However, both techniquescannot assess plaque composition and vulnerability and, therefore,give no information on whether an immediate surgical intervention isimperative or not. Clinically important for an early selection andadministration of an appropriate therapy would be the distinctionbetween those atherosclerotic plaques that are at risk for rupture fromthose that are stable and do not require immediate intervention.

Much effort is put into the search for suitable radiotracers that allownon-invasive imaging of rupture-prone plaques by positron emissiontomography (PET) or single photon emission computed tomography

xmetalloproteinases in human atherosclerotic plaques and evaluationMed Biol (2014), http://dx.doi.org/10.1016/j.nucmedbio.2014.04.085

http://dx.doi.org/10.1016/j.nucmedbio.2014.04.085

mailto:[email protected]



http://www.sciencedirect.com/science/journal/09698051


2 A. Müller et al. / Nuclear Medicine and Biology xxx (2014) xxx–xxx

(SPECT) [5–7]. Recent advances in molecular imaging indicate that thekey to distinguish vulnerable from stable plaques lies in immuneprocesses. Among the most promising PET probes so far are thosetargeting activated immune cells. They include 18 F-fluorodeoxyglucose(18 F-FDG) [8,9], 18 F-fluorodeoxymannose (18 F-FDM) [10] and folate-derived radiopharmaceuticals [11–13], targeting activated macro-phages. We have recently evaluated in human CEA samples the folatereceptor 2 and immune-stimulatory molecule CD80 as promisingimaging targets [7,13].

The high load of atherosclerotic plaques with activated macro-phages provides a prominent source of matrix metalloproteinases(MMP). MMPs play a central role in degradation of extracellularmatrix (ECM), resulting in destabilization of the atheroscleroticplaque whichmay contribute to plaque rupture [14–18]. The integrityof the ECM depends on the balance of degradation and repair ofcollagen and other matrix components [19,20]. Macrophages andmast cells locally destabilize the plaque due to their ability to produceand secrete MMPs. The advantage of imaging active MMPs non-invasively by PET or SPECT could be the assessment of tissuedegradation levels and thus allow detecting plaques in an advancedstate of atherosclerotic lesion when they become harmful.

MMPs are a group of more than 20 structurally related proteasesthat share a zinc-based catalytic mechanism [21]. MMPs aretransferred to the ECM as soluble or membrane-bound inactivezymogens (pro-MMPs). Once activated via the “cysteine switch”mechanism an active site exposing a zinc ion is provided that initiatesits proteolytic activity. In tissue, MMP activity is counterbalanced bythe tissue inhibitors of MMPs (TIMPs), which prevent excessiveproteolytic activity. Among themanyMMPs that have been identified,the two gelatinases MMP-2 (gelatinase A) and MMP-9 (gelatinase B)contribute substantially to the degradation of type IV collagen andgelatin, the two main components of ECM. MMP-2 and MMP-9 areinvolved in a variety of diseases, e.g. cancer, Alzheimer’s disease,and myocarditis [22]. In the context of atherosclerosis, recentpublications reported higher MMP-9 levels in macrophage-richlesions and higher MMP-2 levels in the more stable smooth musclecell-rich lesions [23,24].

Several research groups are currently working on the developmentand evaluation of MMP inhibitor-based radiotracers targeting MMPsin their active form [25–28]. MMP inhibitors have been evaluated asdiagnostic and therapeutic tools in cancer, autoimmune disease, andcardiovascular disease [29,30]. Despite the over 3000 entries under“MMP inhibitors” in the data base Integrity (Thomson Reuters), ofwhichmany have been tested both experimentally and clinically, onlyone or two, including the antibiotic doxycycline, are approved forsystemic use as MMP inhibitor by the Food and Drug Administration.The low success rate is mainly due to the undesirable side effects ofMMP inhibitors in particular on the musculoskeletal system. Most arebroad spectrum MMP inhibitors or at least not selective for onespecific MMP-type. This is not only a drawback for therapy but is alsoa limitation for the imaging approach with respect to the specificity ofthe resulting signal.

We have recently introduced the radiosynthesis of a selective lowmolecular weight MMP-9/MMP-2 inhibitor as a potential PET probefor the two gelatinases [27]. In the present study, excised humanartery wall segments from CEA classified into normal arteries, stableand vulnerable atherosclerotic plaques were investigated regardingtheir gene expression levels of various MMPs. We further character-ized this tissue regarding the catalytically active forms of MMP-2 andMMP-9, the targets of our radiotracer [27]. We tested in vitro theability of the tracer as an imaging agent for atherosclerotic plaquesand further evaluated the 3H-labeled selective MMP-9 inhibitormethyl 4-[3-(formylhydroxyamino)-4-(4-(4-trifluoromethoxy-phenoxy)phenylsulfonyl)butyl]benzoate (3H-2, compound 19 g inreference [31]) as a potential probe to specifically target activeMMP-9in vulnerable atherosclerotic plaques.

Please cite this article as: Müller A., et al, Gene expression levels ofmatriof radiolabeled inhibitors as imaging agents for plaque vulnerability, Nucl

2. Material and methods

2.1. Clinical data and human tissue banking

Patients were referred to the Clinic for Cardiovascular Surgery ofthe University Hospital of Zurich (Zurich, Switzerland) for CEAbetween 2010 and 2012. Written informed consent was obtainedfrom all patients. Before surgery, all patients underwent Duplexultrasound (to define the degree of stenosis and the morphologiccharacteristics of plaques) or computed tomography angiography,as well as a computed tomography or magnetic resonance brainscan. The atherosclerotic plaques were removed from the internal,external and common carotid artery by bifurcation advancement[32]. All excised specimens were photographed for macroscopic re-evaluation, transferred within 5 min into RNAlater® solution(Sigma, St. Louis, MO, USA), stored at 4 °C overnight and at−80 °Cuntil further use.

2.2. RNA isolation and quantitative polymerase chain reaction

Total RNA was extracted from plaque specimens using the Isol-RNA Lysis Reagent (5 Prime, Gaithersburg, USA) and the bead-milling TissueLyser system (Qiagen, Hilden, Germany). QuantiTect®Reverse Transcription Kit (Qiagen, Hilden, Germany) was used togenerate cDNA. Primers were synthesized by Microsynth (Balgach,Switzerland) or purchased from Qiagen (Hilden, Germany). Primersequences are listed in Supplemental Table 1. Quantitative poly-merase chain reaction (qPCR) was performed with the DyNAmoTMFlash SYBR® Green qPCR Kit (Thermo Scientific) using a 7900 HTFast Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA).Beta-actin (ACTB) primers were used as internal control to assessamplification efficiency. The amplification signals were detected inreal-time, which permitted accurate quantification of the amountsof the initial RNA template during 50 cycles. All reactions wereperformed in duplicates and in three independent runs. Quantita-tive analysis was performed using the SDS software v2.4 (AppliedBiosystems, Carlsbad, CA, USA) and a previously described quanti-fication method [33]. The specificity of the PCR products of each runwas determined and verified with SDS Software's dissociation curveanalysis feature.

2.3. Immunohistochemical stainings

For histology and immunohistochemistry, plaques were embed-ded in Neg-50 Frozen Section Medium (Thermo Scientific, Runcorn,UK) and cut in 7 μm serial slices on a cryostat (Cryo-Star HM 560MV; Microm International GmbH), thaw-mounted onto coated glassslides (Superfrost Plus, Thermo Scientific, Runcorn, UK), and storedat −80 °C until further use. For histological characterization of theplaque morphology hematoxylin and eosin (HE; (Sigma, St. Louis,MO, USA) staining was applied.

Immunohistochemistry was performed according to the protocolof Dako REALTM EnVisionTM Detection System, Peroxidase/DAB,Rabbit/Mouse (Dako, Carpinteria, CA, USA) using mouse anti-CD68(1:50) (ab955, Abcam, Cambridge, UK), rabbit anti-MMP-2 (1:100)(ab79781, Abcam, Cambridge, UK) or mouse anti-MMP-9 (1:300)(ab3195, Abcam, Cambridge, UK) antibodies. Sections were counter-stained with hematoxylin solution, Gill No. 1 (Sigma, St. Louis, MO,USA). Corresponding immunoglobulin G (IgG) antibodies served asinternal negative control.

All sections were covered using Eukitt mounting medium(Sigma) and analyzed under a microscope (Zeiss AxioImager A1,Jena, Germany or digital slide scanner Pannoramic 250, Sysmex,Horgen, Switzerland).



Table 1Characteristics and clinical data of the patients included in this study.

Asymptomaticpatients

Symptomaticpatients

Number of patients 5 10Numbers of female/male 1/4 4/6Age at surgery (y), mean ± S.D. 69.4 ± 3.6 71.3 ± 6.1Diabetes mellitus (%) 60 60Arterial hypertension (%) 40 70Hypercholesterolemia (%) 80 70Smoking (%) 80 60Indication for CEAStroke (%) 0 60Transient ischemic attack (%) 0 40On-pumpsurgery topreventperioperative stroke (%) 20 0Ultrasound revealed instable and soft plaques (%) 20 0Others: aortic valve or coronary artery surgery (%) 60 0Number of normal arteries 4 2Number of stable plaques 5 10Number of vulnerable plaques 3 8

3A. Müller et al. / Nuclear Medicine and Biology xxx (2014) xxx–xxx

2.4. Protein extraction and Western blotting

Human plaque segments and mouse lung tissue were disruptedand homogenized in liquid nitrogen using mortar and pestle. Thistissue powder or NCI-H69 cells (small cell lung cancer cells, DMSZ,Braunschweig, Germany) were harvested in CelLytic™ MT Cell LysisReagent (Sigma, St. Louis, MO, USA) containing Halt Protease InhibitorCocktail (Thermo Scientific, Runcorn, UK). Tissue lysates wereequilibrated on ice for 30 min, followed by ultrasound disruptionand centrifugation at 12,000 g for 10 min, 4 °C. Proteins of thesupernatants were separated by 10% sodium-dodecyl-sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE), according to standardprotocols (Bio-Rad, Hercules, MA, USA). After SDS-PAGE, proteinswere transferred to nitrocellulose membranes (GE Healthcare, LittleChalfont, UK). The blots were blocked in 5% dried milk in TRIS-bufferedsaline-Tween-20 (TBS-T) and processed for immunodetection withpolyclonal anti-MMP-2 (1:500) (ab79781, Abcam, Cambridge, UK),anti-MMP-9 (1:500) (ab38898, Abcam, Cambridge, UK) or anti-β-actin(A2066, Sigma, St. Louis, MO, USA) antibody at 4 °C overnight. Boundantibodies were visualized with anti-rabbit IgG secondary antibodiesconjugated with horseradish peroxidase diluted at 1:80,000 (Sigma) byenhanced chemiluminescence.

2.5. In vitro autoradiography with the 18F-labeled MMP-2/MMP-9inhibitor (18F-1)

In vitro autoradiography experiments were performedwith 20 μmthick cryo-sections mounted on glass slides. Slices were allowed tothaw at room temperature (rt) prior to the experiments. Forautoradiography with our previously reported 18F-labeled MMP-2/MMP-9 ligand (R)-2-(4-(4-fluorobenzamido)phenylsulfonamido)-3-(1H-indol-3-yl)propanoic acid [27] (18F-1, Supplementary Fig. 2)slices were washed in TRIS buffer (50 mM TRIS/HCl, pH 7.4)containing 0.1% bovine serum albumin (TRIS/BSA) for 10 min at4 °C. Slices were incubated with 1 nM 18F-1 in TRIS/BSA for 1 hour ina humified chamber at rt. Nonspecific binding was determined inthe presence of 10 μM unlabeled 1. The slices were washed 3 ×5 min in TRIS/BSA, 2 × 3 min in TRIS buffer and 2 × 5 s in distilledwater at 4 °C. The air-dried slices were exposed to a BAS-MS 2025phosphor imaging plate (Fuji Film, Dielsdorf, Switzerland) for60 min, scanned in a BAS-5000 bio-imaging reader (Fuji Film,Dielsdorf, Switzerland) and analyzed with the AIDA 4.5 software(Raytest, Sprockhövel, Germany).

2.6. In vitro autoradiography with selective MMP-9 inhibitor 3H-2

Autoradiography was in addition performed with the 3H-labeledselective MMP-9 ligand 3H-2 (Supplemental Fig. 2). This compoundwas tritiated by iridium catalyzed H-3H exchange reaction and kindlyprovided by Bayer Healthcare (Berlin, Germany). Synthesis of theunlabeled molecule has been reported elsewhere [31]. The plaqueslices were pre-incubated for 10 min in advanced TRIS buffer (50 mMTRIS/HCl pH 7.4, 150 mMNaCl, 10 mM CaCl2, 0.02% NaN3, 0.05% (v/v)Brij®35) with 0.1% BSA at 4 °C. To inactivate MMPs, the chelatorEDTA, EGTA or tetracycline was added to the pre-incubation buffer atthe indicated concentration. Afterwards, slices were incubated with100 μL of 3H-2 solution (1 nM) in the respective advanced TRIS bufferwith or without chelator for 1 hour at rt in a humid chamber. Afterdecanting, the slices were washed and dried as described above. Sliceswere exposed to a BAS-TR 2025 phosphor imaging plate (Fuji Film,Dielsdorf, Switzerland) for 1–2 weeks, scanned and analyzed asdescribed above.

The amount of radioactivity bound to the plaque samples wasquantified with a calibration curve derived from tritium standards ona glass slide (range: 0–489.1 nCi/mg, ARC, St. Louis, MO, USA) andexposed to the same plate. Radioactivity signalswerewithin the linear

Please cite this article as: Müller A., et al, Gene expression levels of matriof radiolabeled inhibitors as imaging agents for plaque vulnerability, Nucl

range of the calibration curve. Tissue radioactivity, normalized toplaque size, was corrected for the accumulated radioactivity afterMMP inactivation (2 mM EDTA) to calculate the specific binding. Thespecific binding was normalized to one sample of the normal arteries(value of specific binding equals 1 for this sample) serving asreference in all experiments and, therefore, expressed as relativebinding. In the end, all sections were HE-stained as described above.Experiments were done three times.

2.7. Statistics

Statistical analysis was performed using Sigma Plot 12.5software. Quantitative PCR data were analyzed by one way analysisof variance (ANOVA) tests, followed by corrections for multiplepost-hoc tests (Holm-Sidak, Bonferroni). Binding levels of in vitroautoradiography experiments were compared using ANOVA test,followed by the Holm-Sidak post-hoc test. P b 0.05 was consideredas statistically significant.

3. Results

3.1. Clinical data and classification criteria

A total of 15 patients were included in this study, ranging in agefrom 62 to 81 years. Patients were divided into an asymptomatic anda symptomatic group. Further patients’ characteristics and clinicaldata are summarized in Table 1.

Plaques were classified, according to the American Heart Associ-ation guidelines, as stable if they macroscopically presented a lipidcore encapsulated between a fibrous cap and the media withoutendothelial disruption. Plaques with intraplaque hemorrhage evolv-ing in the necrotic core, ulcerated endothelial surface or thrombosiswere considered vulnerable [34]. This macroscopic classification wasconfirmed by histological analysis with HE-staining (SupplementalFig. 1) and as described previously [13,35]. Human tissue samples of 6normal artery walls, 15 stable and 11 specimens of vulnerable plaqueswere used. The normal artery walls were redundant segments, mostlyfrom Arteria thyroidea superior or Arteria sternocleidomastoidea whichhad to be removed during surgeries.

3.2. MMP, TIMP-3 and CD68 gene expression levels and correlations

Members of the following MMP families were investigated fortheir messenger RNA levels in human normal arteries, stable andvulnerable plaques. From the collagenase family: MMP-1, MMP-8;gelatinase family: MMP-2, MMP-9; stromelysins: MMP-10; mem-




brane-type MMP family: MMP-14; and MMP-12 representing ametalloelastase. All screened MMPs were significantly upregulatedin both plaque types compared to normal artery walls (Fig. 1a).Among them, the expression of MMP-1 (1.5-fold; P = 0.02), MMP-9(5.2-fold; P b 0.001), MMP-12 (1.4-fold; P = 0.005) and MMP-14(1.5-fold; P = 0.016) was significantly higher in vulnerable thanstable plaques. The level of TIMP-3, a potent inhibitor of MMP-9 [36],

Fig. 1. Transcription levels of different MMPs, TIMP-3 and CD68. a. Relative expression of MMarteries (open circles), stable plaques (gray circles), and vulnerable plaques (black circlesstatistical values. The boundaries of the box indicate the 25th and the 75th percentile, respecabove and below the box indicate the 10th and 90th percentiles. Outliers are indicated byMMP-2 and MMP-9 with TIMP-3 or CD68. Open circles, normal arteries; gray circles, stable


was significantly higher in stable plaque types than normal arteries(8.9-fold; P = 0.002) and significant lower in vulnerable compared tostable plaques (4-fold; P = 0.002). Expression of TIMP-3 was ofsimilar low level in vulnerable plaques and normal arteries. Fig. 1bshows, that TIMP-3 weakly correlated positively with MMP-2 (Rall =0.46 including all data points; RPlaques = 0.26 excluding data point fromnormal arteries) but negatively with MMP-9 levels (Rall = −0.20;

P-1, MMP-2, MMP-8, MMP-9, MMP-10, MMP-12, MMP-14, TIMP-3 and CD68 in normal). Expression levels were calculated relative to β-actin (n = 3). Box plots representtively. A dashed line graphs the mean and a solid line the median. Whiskers (error bars)black dots outside the whiskers. b. Correlation between relative transcription levels ofplaques; black circles, vulnerable plaques.

x metalloproteinases in human atherosclerotic plaques and evaluationMed Biol (2014), http://dx.doi.org/10.1016/j.nucmedbio.2014.04.085


Fig. 2. Identification of MMP-2 and MMP-9 protein in vulnerable and stable plaquesRepresentative immunohistochemical stainings of a normal artery (a, b), stable (c–f)and vulnerable (g–j) plaques using anti-CD68, anti-MMP-2, and anti-MMP-9 antibodiesor control IgG (omitting primary antibody). Boxes in a, c, e, g and i indicate high-magnification areas shown below each plaque in b, d, f, h and j. Boxes apply for all fourstainings. Sections were counterstained with hematoxylin, resulting in blue nucleiScale bar: a: 1 mm; c, e, g, i: 2 mm; b, d, f, h, j: 100 μm. (K) Western blot analysis oactive forms of MMP-2 and MMP-9 in normal arteries, stable and vulnerable plaques(different tissue samples than in a–j), and NCI-H69 cells as positive control (P) forMMP-2 expression or murine lung for MMP-9.


RPlaques = −0.54). Correlation of TIMP-3 and MMP-9 expressionrevealed distinct clusters of data points from normal arteries, stable orvulnerable plaques. MMP-2 and MMP-9 expression weakly correlated(Rall = 0.26; RPlaques = −0.01).

RNA levels of the macrophage marker CD68 were on average 39-fold higher in atherosclerotic plaques than normal arteries(P b 0.05) with no significant difference between both plaquetypes. The expression of CD68 had a moderate correlation withMMP-2 (Rall = 0.24; RPlaques = −0.13) and MMP-9 (Rall = 0.36;RPlaques = 0.21).

3.3. Localization and correlation of CD68-positive macrophages withMMP-2 and MMP-9 and verification of their active forms

To localize MMP-2 and MMP-9 expressing cells within the plaquesin relation to CD68-positive macrophages serial tissue sections wereimmunostained accordingly (Fig. 2a–j). Non-immune IgG controlsrevealed that the staining was specific. Normal artery walls lackedpositive staining for CD68, MMP-2 and MMP-9 (Fig. 2a, b). Stableplaques presented in their lipid core a positive staining for CD68,MMP-2 and weak MMP-9 intensities (Fig. 2c–f). Vulnerable plaquesshowed strong staining for CD68, MMP-2 and MMP-9 (Fig. 2 g–j).

The majority of MMP-2- and MMP-9-positive cells were located inthe deeper part of the plaque surrounding the necrotic core and in thecap region. Distribution patterns of the two MMPs overlapped in thevulnerable plaques but were not identical. Most, but not all CD68-positive cells stained positive for MMP-2 or MMP-9.

MMP-2 and MMP-9 protein expression was analyzed by Westernblot in 5 normal arteries, 10 stable plaques and 8 vulnerable plaques. Arepresentative Western blot is shown in Fig. 2 k. The polyclonalantibody anti-MMP-2 recognized the latent pro-enzyme at 72 kDaand the enzymatically active form at 67 kDa. Human small cell lungcancer cells (NCI-H69) served as positive control (indicated with P inFig. 2 k) [37]. Pro- and active MMP-2 was detected in stable as well asvulnerable plaques, but not in normal arteries. The pro-form of MMP-9 was detected at 92 kDa and the active forms at 82, 67 and 64 kDa.Mouse lung tissue was positive for both MMP-9 forms with bandsappearing at higher molecular weights than observed for the humanortholog because of the longer peptide sequence of mouse MMP-9[38]. Normal arteries revealed only pro-MMP-9 whereas stable andvulnerable plaques additionally possessed the enzymatically activeform of MMP-9.

3.4. In vitro autoradiography with 18F-MMP-2/MMP-9 inhibitor 1 (18F-1)

The high binding affinity (IC50[MMP-2] = 1.8 nM; IC50[MMP-9] = 7 nM) and in vivo stability of 18F-1 that we reported recently[27] encouraged us to evaluate this compound further and perform invitro autoradiography studies on human atherosclerotic plaque slices.The gelatinase inhibitor 18F-1 bound preferentially to atheroscleroticplaques with hot-spot binding to areas with intimal lesions. Fig. 3shows one representative sample each of 2 normal arteries, 4 stableplaques and 4 vulnerable plaques. The radioactivity signal wasindependent of the plaque type. Addition of excess unlabeled 1confirmed specific binding as it reduced the signal in the lesions, albeitnot to the level of healthy artery wall.

3.5. In vitro autoradiography with 3H-MMP-9 inhibitor (3H-2) underbaseline and MMP inactivating conditions

As shown recently [23,24] and confirmed by qPCR in this work,MMP-9 is associated with plaque vulnerability while MMP-2 is higherin stable plaques. We performed autoradiography with 3H-2, whichhas a nearly 35-fold selectivity for MMP-9 over MMP-2. Wada et al.[31] reported an IC50 of 0.98 nM for MMP-9 and 34 ± 5.1 nM forMMP-2. Representative autoradiograms showed marginal binding

Please cite this article as: Müller A., et al, Gene expression levels of matrixmetalloproteinases in human atherosclerotic plaques and evaluationof radiolabeled inhibitors as imaging agents for plaque vulnerability, Nucl Med Biol (2014), http://dx.doi.org/10.1016/j.nucmedbio.2014.04.085

.

.f

of 3H-2 to normal arteries but increased binding to both plaquetypes (Fig. 4a). The pattern of radioactivity was in agreement withthe immunohistochemical staining patterns for MMP-9 on adjacentslices. Binding of 3H-2 was 2-fold higher on plaque tissue (n = 18)than normal arteries (n = 5; P = 0.005) but was similar invulnerable (n = 7) and stable (n = 11) plaques with a ratio of1.2 (P = 0.137; Fig. 4b).


Fig. 3. Representative autoradiography images of a human normal artery, stable and vulnerable plaque using the MMP-2/MMP-9 inhibitor 18F-1. For blocking conditions excess ofunlabeled 1 was added. Autoradiograms are superimposed onto morphological hematoxylin-eosin (HE)-stained slices. Color scale for low to high 18F-1 binding. Scale bar for low-magnification images: 2 mm, and high-magnification images in bottom row: 500 μm.


Themetal chelators EDTA and EGDT as well as the broad-spectrumantibiotic tetracycline are inhibitors of the gelatinolytic activity ofMMP-9 [39]. To test whether 3H-2 binds to the active form of MMPswe pre-incubated the plaque slices with either of these zinc chelatorsbefore adding the tritiated ligand. The 3H-2 bindingwas diminished ina dose-dependent manner with the highest effect at 5 mM chelator.Fig. 4 shows the effects of EDTA. EGTA and tetracycline revealedsimilar results (data not shown).

4. Discussion

Our qPCR and IHC analyses confirmed the high load with CD68-rich cells from the macrophage lineage in atherosclerotic plaques. Asexpected for macrophage-infiltrated tissues, MMP mRNA levelswere increased in the atherosclerotic lesions. Of the investigatedMMPs, MMP-1, MMP-9, MMP-12 and MMP-14 were significantlyhigher in vulnerable than stable plaques with the highest ratioobserved for MMP-9.

Initiation of collagen breakdown in plaques requires the interstitialcollagenases MMP-1 andMMP-8. In our study, expression of both wassignificantly up-regulated in atherosclerotic plaques with a tendencyto higher levels (P b 0.05 for MMP-1) in vulnerable plaques. This is inagreement with previous findings. Sukhova and coworkers haveshown that the increased collagenolysis in atheromatous plaques ismediated by MMP-1 [40]. In a cohort of 159 patients, theconcentration of active MMP-8 was significantly increased in carotidplaques with vulnerable appearance and MMP-8 protein co-localizedwith macrophage-positive areas [41]. Knocking out MMP-8 substan-tially reduced the extent of atherosclerotic lesions inWestern diet-fedApoE knockout mice, a model of atherosclerosis [42].


Besides the collagenases, the gelatinases MMP-2 and in particularMMP-9 gained much attention in atherosclerosis research [23,24,43].As described in the introduction, MMP-2 is involved in plaquestabilization while MMP-9 activity is associated with plaque desta-bilization and disruption [6]. MMP-9 promotes matrix degradationfacilitating the entry of immune cells into lesioned sites andneovascularization. We confirmed this distribution pattern inhuman carotid lesions on the mRNA level. RNA levels of MMP-2were increased in stable lesions compared to normal artery walls, butremained unaltered in vulnerable plaques. The active form of MMP-2was present in both plaque types. MMP-9 gene expression wassignificantly increased in vulnerable plaques compared to stablephenotypes, and active MMP-9 protein was present in both asconcluded from Western blot analysis. As MMP-9 is a key player inplaque destabilization, one may hypothesize that enhanced MMP-9activity in stable plaques may be involved in the transition to a morevulnerable type.

MMP-12 and MMP-14 levels were significantly higher in vulner-able plaques underlying their pro-inflammatory and detrimental role.MMP-14 deficiency in bone marrow-derived cells leads to collagenaccumulation in murine atherosclerotic plaques, indicating therelevance of MMP-14 in the pathophysiology of atherosclerosis [44].However, loss of MMP-12 in mice showed no clear evidence for asignificant involvement in atherosclerotic plaque growth [43].

Besides MMP-2 and MMP-9 we did not proof the presence of theenzymatically active forms of the other MMPs, but our measured geneexpression levels correlated well with the above described observa-tions by others. Further publications demonstrated, that MMP geneexpression is strictly regulated in progressing atherosclerosis bycytokines and growth factors [45,46]. Expression of TIMP-3 was highin stable plaques and low with almost status of normal arteries, in



Fig. 4. a. Representative images of MMP-9 immunostaining, in vitro 3H-2 autoradiog-raphy and hematoxylin-eosin (HE) staining. Adjacent sections were incubated inincreasing concentration of EDTA (0, 2, 5 mM). Gray scale for low to high 3H-2 binding.Scale bars: 2 mm. b. Semiquantitative autoradiography-derived measurement for thespecific binding of 3H-2 to normal arteries, stable and vulnerable plaques.


vulnerable plaques. Increased ratios of MMP-9 to TIMPs were shownfor disrupted plaques [6], with similar TIMP-3 expression levels indisrupted plaques as in control tissue, in agreement with our data forvulnerable plaques. The increased TIMP-3 expression that weobserved in stable plaques is interesting in the context of its anti-inflammatory function [5]. TIMP-3 is an irreversible inhibitor ofseveral MMPs, including MMP-2 and MMP-9 [47], and a more potentinhibitor of MMP-9 than other TIMPs [36]. TIMP-3 may thus beinvolved in the stabilization of atherosclerotic lesions by inhibitinginflammatory and ECM-degrading processes. It has in addition beenreported as key regulator with protective function in the process ofmyocardial failure [47].

According to our own findings and reports by others cited above,active MMP-9 could serve as a good target for the non-invasiveimaging of plaque vulnerability. Imaging MMP activity could inaddition be of high interest in other inflammatory diseases andoncology. Our 18F-labelled MMP-2/MMP-9 inhibitor showed higheraccumulation in human atherosclerotic plaques than in healthy arterywalls in vitro. However 18F-1 accumulated in both plaque types. Thiswas not unexpected from its similar affinities for the two gelatinases

Please cite this article as: Müller A., et al, Gene expression levels of matriof radiolabeled inhibitors as imaging agents for plaque vulnerability, Nucl

and our gene expression and immunohistochemistry results showinghigh MMP-9 in vulnerable and high MMP-2 in stable plaques.

Aiming to discriminate atherosclerotic lesions at higher risk forrupture from stable lesions, the potential of the MMP-9 selectiveradiotracer 3H-2was evaluated.We confirmed its selectivity for activeover inactive MMP in the autoradiography experiment by addition ofmetal chelators trapping the Zn2+-ion from the active site silencingMMP activity. Binding of 3H-2 markedly detected carotid plaques invitro. Despite higher MMP-9 expression levels in vulnerable plaques,it was not able to significantly discriminate stable from vulnerablelesions. We can only speculate why accumulation was similar in bothplaque types despite the higher MMP-9 expression levels invulnerable plaques. MMP-9 activity may be similar in both plaquetypes despite the difference in expression level. This hypothesiswould be supported by our Western blot analysis, showing bands ofactiveMMP-9 in both vulnerable and stable plaques. Alternatively, theinhibitor may bind to one or more additional proteases which areincreased in stable plaques. Selectivity was shown for MMP-9 overMMP-2 and MMP-1 while other proteases were not tested. Inaddition, MMP-9/MMP-2 selectivity of 3H-2 is relatively low with anaffinity ratio of 35 [31]. This would not allow distinguishing MMP-9fromMMP-2 rich regions in cases where MMP-2 levels are higher by asimilar or greater factor. Based on our data, we cannot conclude on theexpression or activity ratio between the two enzymes. A thirdpossibility is an increase in MMP-2 affinity by hydrolysis of thecompound by tissue esterases to its respective benzoate, whichshowed similar affinity for the two gelatinases [31]. Despite greatefforts in this field, we are not aware of any other low molecularweight compound targeting active MMP with high selectivity forMMP-9 [31,48]. Based on our results and reports from others we areconvinced that a low molecular weight compound targeting selec-tively active MMP-9 could become the lead structure of a successfulimaging agent not only for atherosclerosis but also for otherinflammatory diseases or tumors.

This study supports the leading role of MMPs, in particular MMP-1,MMP-9, MMP-12 and MMP-14 in advanced carotid plaques. Furtherefforts in medicinal chemistry may bring light to binders whicheventually allow selective targeting of these MMPs. Imaging-baseddiagnostic tools could emerge from these compounds to detectrupture prone plaques.

In conclusion, we have shown increased expression levels of MMP-1, -9, -12 and -14 in vulnerable and MMP-2 and TIMP-3 in stableplaques. Results for MMP-9 and -2 were confirmed by IHC.Radiolabeled MMP-2/-9 inhibitors were able to discriminate in invitro autoradiography experiments between atherosclerotic lesionsand healthy artery walls. We anticipate that non-invasive imaging ofvulnerable plaques will be feasible with a tracer of high selectivity foractive MMP-9.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.nucmedbio.2014.04.085.

Conflict of interest(s)/disclosure(s)

None.

Acknowledgments

Part of this work was supported by the Clinical Research PriorityProgram of the University of Zurich onMolecular Imaging (MINZ).Weare grateful to Petra Wirth for technical support. We thank Jelena-Rima Ghadri, Michael Fiechter, Tobias Fuchs and Julia Stehli forcoordinating the constant collection of human tissue samples andMarkus Schmid for taking care of the standardized process of tissuecollection in the operating room of the University Hospital Zurich,Switzerland. The authors acknowledge support from the Scientific






Center for Optical and Electron Microscopy ScopeM of the SwissFederal Institute of Technology ETHZ.

References

[1] Rosamond W, Flegal K, Friday G, Furie K, Go A, Greenlund K, et al. Heart diseaseand stroke statistics–2007 update: a report from the American Heart AssociationStatistics Committee and Stroke Statistics Subcommittee. Circulation 2007;115:e69-171.

[2] Liapis CD, Bell PR, Mikhailidis D, Sivenius J, Nicolaides A. Fernandes e Fernandes J,et al. ESVS guidelines. Invasive treatment for carotid stenosis: indications,techniques. Eur J Vasc Endovasc Surg 2009;37:1–19.

[3] Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl JMed 2005;352:1685–95.

[4] Libby P. Inflammation in atherosclerosis. Arterioscler Thromb Vasc Biol2012;32:2045–51.

[5] Khokha R, Murthy A, Weiss A. Metalloproteinases and their natural inhibitors ininflammation and immunity. Nat Rev Immunol 2013;13:649–65.

[6] Higashikata T, Yamagishi M, Higashi T, Nagata I, Iihara K, Miyamoto S, et al. Alteredexpression balance of matrix metalloproteinases and their inhibitors in humancarotid plaque disruption: results of quantitative tissue analysis using real-timeRT-PCR method. Atherosclerosis 2006;185:165–72.

[7] Müller A, Mu L, Meletta R, Beck K, Rancic Z, Drandarov K, et al. Towards non-invasiveimaging of vulnerable atherosclerotic plaques by targeting co-stimulatory molecules.Int J Cardiol 2014. http://dx.doi.org/10.1016/j.ijcard.2014.04.071.

[8] Rudd JH, Warburton EA, Fryer TD, Jones HA, Clark JC, Antoun N, et al. Imagingatherosclerotic plaque inflammation with [18 F]-fluorodeoxyglucose positronemission tomography. Circulation 2002;105:2708–11.

[9] Davies JR, Rudd JH, Fryer TD, Graves MJ, Clark JC, Kirkpatrick PJ, et al. Identificationof culprit lesions after transient ischemic attack by combined 18 F fluorodeox-yglucose positron-emission tomography and high-resolution magnetic resonanceimaging. Stroke 2005;36:2642–7.

[10] Tahara N, Mukherjee J, de Haas HJ, Petrov AD, Tawakol A, Haider N, et al. 2-deoxy-2-[F]fluoro-d-mannose positron emission tomography imaging in atherosclerosis.Nat Med 2014. http://dx.doi.org/10.1038/nm.3437.

[11] Ayala-Lopez W, Xia W, Varghese B, Low PS. Imaging of atherosclerosis inapoliprotein e knockout mice: targeting of a folate-conjugated radiopharmaceu-tical to activated macrophages. J Nucl Med 2010;51:768–74.

[12] Winkel LC, Groen HC, van Thiel BS, Müller C, van der Steen AF, Wentzel JJ, et al.Folate receptor-targeted single-photon emission computed tomography/computed tomography to detect activated macrophages in atherosclerosis:Can It distinguish vulnerable from stable atherosclerotic plaques? MolImaging 2013;12:1–5.

[13] Müller A, Beck K, Rancic Z, Müller C, Fischer CR, Betzel T, et al. Imagingatherosclerotic plaque inflammation via folate receptor targeting using a novel18 F-folate radiotracer. Mol Imaging 2014;13:1–11.

[14] Fujimoto S, Hartung D, Ohshima S, Edwards DS, Zhou J, Yalamanchili P, et al.Molecular imaging of matrix metalloproteinase in atherosclerotic lesions:resolution with dietary modification and statin therapy. J Am Coll Cardiol2008;52:1847–57.

[15] Hermann S, Starsichova A, Waschkau B, Kuhlmann M,Wenning C, Schober O, et al.Non-FDG imaging of atherosclerosis: will imaging of MMPs assess plaquevulnerability? J Nucl Cardiol 2012;19:609–17.

[16] Sadeghi MM, Glover DK, Lanza GM, Fayad ZA, Johnson LL. Imaging atherosclerosisand vulnerable plaque. J Nucl Med 2010;51(Suppl 1):51S–65S.

[17] Schäfers M, Schober O, Hermann S. Matrix-metalloproteinases as imaging targetsfor inflammatory activity in atherosclerotic plaques. J Nucl Med 2010;51:663–6.

[18] Newby AC. Do metalloproteinases destabilize vulnerable atherosclerotic plaques?Curr Opin Lipidol 2006;17:556–61.

[19] Galis ZS, Sukhova GK, Lark MW, Libby P. Increased expression of matrixmetalloproteinases and matrix degrading activity in vulnerable regions ofhuman atherosclerotic plaques. J Clin Invest 1994;94:2493–503.

[20] Pepper MS. Extracellular proteolysis and angiogenesis. Thromb Haemost2001;86:346–55.

[21] Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metallopro-teinases: structure, function, and biochemistry. Circ Res 2003;92:827–39.

[22] Malemud CJ. Matrix metalloproteinases (MMPs) in health and disease: anoverview. Front Biosci 2006;11:1696–701.

[23] Sluijter JP, Pulskens WP, Schoneveld AH, Velema E, Strijder CF, Moll F, et al. Matrixmetalloproteinase 2 is associated with stable andmatrix metalloproteinases 8 and9 with vulnerable carotid atherosclerotic lesions: a study in human endarterec-


tomy specimen pointing to a role for different extracellular matrix metallopro-teinase inducer glycosylation forms. Stroke 2006;37:235–9.

[24] Hobeika MJ, Thompson RW, Muhs BE, Brooks PC, Gagne PJ. Matrix metallopro-teinases in peripheral vascular disease. J Vasc Surg 2007;45:849–57.

[25] Hugenberg V, Breyholz HJ, Riemann B, Hermann S, Schober O, Schäfers M, et al. Anew class of highly potent matrix metalloproteinase inhibitors based on triazole-substituted hydroxamates: (radio)synthesis and in vitro and first in vivoevaluation. J Med Chem 2012;55:4714–27.

[26] Schrigten D, Breyholz HJ, Wagner S, Hermann S, Schober O, Schäfers M, et al. Anew generation of radiofluorinated pyrimidine-2,4,6-triones as MMP-targetedradiotracers for positron emission tomography. J Med Chem 2012;55:223–32.

[27] Selivanova SV, Stellfeld T, Heinrich TK, Müller A, Krämer SD, Schubiger PA, et al.Design, synthesis, and initial evaluation of a high affinity positron emissiontomography probe for imaging matrix metalloproteinases 2 and 9. J Med Chem2013;56:4912–20.

[28] Wagner S, Faust A, Breyholz HJ, Schober O, Schäfers M, Kopka K. TheMMP inhibitor (R)-2-(N-benzyl-4-(2-[18 F]fluoroethoxy)phenylsulphonamido)-N-hydroxy-3-methylbuta namide: Improved precursor synthesis and fullyautomated radiosynthesis. Appl Radiat Isot 2011;69:862–8.

[29] Benjamin MM, Khalil RA. Matrix metalloproteinase inhibitors as investigativetools in the pathogenesis and management of vascular disease. EXS2012;103:209–79.

[30] Scherer RL, McIntyre JO, Matrisian LM. Imaging matrix metalloproteinases incancer. Cancer Metastasis Rev 2008;27:679–90.

[31] Wada CK, Holms JH, Curtin ML, Dai Y, Florjancic AS, Garland RB, et al.Phenoxyphenyl sulfone N-formylhydroxylamines (retrohydroxamates) as potent,selective, orally bioavailable matrix metalloproteinase inhibitors. J Med Chem2002;45:219–32.

[32] Nett PC, Zund G, Pretre R, Niederhouser U, Vogt PR, Turina M. A 20-year follow-upof internal carotid artery endarterectomy with bifurcation advancement. ThoracCardiovasc Surg 2000;48:279–84.

[33] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-timequantitative PCR and the 2−ΔΔCt Method. Methods 2001;25:402–8.

[34] Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagov S, Insull Jr W, et al. Adefinition of advanced types of atherosclerotic lesions and a histologicalclassification of atherosclerosis. Circulation 1995;92:1355–74.

[35] Müller A, Mu L, Meletta R, Beck K, Rancic Z, Drandarov K, et al. Towards non-invasiveimagingof vulnerable atheroscleroticplaquesby targeting co-stimulatorymolecules. IntJ Cardiol 2014. http://dx.doi.org/10.1016/j.ijcard.2014.04.071 [accepted for publication].

[36] Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behavior.Annu Rev Cell Dev Biol 2001;17:463–516.

[37] Gonzalez-Avila G, Iturria C, Vadillo F, Teran L, Selman M, Perez-Tamayo R. 72-kD(MMP-2) and 92-kD (MMP-9) type IV collagenase production and activity indifferent histologic types of lung cancer cells. Pathobiology 1998;66:5–16.

[38] Sands MF. Localization of matrix metalloproteinase (MMP)-9 in lung tissue of amurine model of allergic asthma. Immunol Invest 2012;41:87–96.

[39] Lee D-H, Kang K. Purification and characterization of human 92-kDa type IVcollagenase (gelatinase B). Exp Mol Med 1996;28:161–5.

[40] Sukhova GK, Schonbeck U, Rabkin E, Schoen FJ, Poole AR, Billinghurst RC, et al.Evidence for increased collagenolysis by interstitial collagenases-1 and -3 invulnerable human atheromatous plaques. Circulation 1999;99:2503–9.

[41] Molloy KJ, ThompsonMM, Jones JL, Schwalbe EC, Bell PR, Naylor AR, et al. Unstablecarotid plaques exhibit raised matrix metalloproteinase-8 activity. Circulation2004;110:337–43.

[42] Laxton RC, Hu Y, Duchene J, Zhang F, Zhang Z, Leung KY, et al. A role of matrixmetalloproteinase-8 in atherosclerosis. Circ Res 2009;105:921–9.

[43] Luttun A, Lutgens E, Manderveld A, Maris K, Collen D, Carmeliet P, et al. Loss ofmatrix metalloproteinase-9 or matrix metalloproteinase-12 protects apolipopro-tein E-deficient mice against atherosclerotic media destruction but differentiallyaffects plaque growth. Circulation 2004;109:1408–14.

[44] Schneider F, Sukhova GK, Aikawa M, Canner J, Gerdes N, Tang SM, et al. Matrix-metalloproteinase-14 deficiency in bone-marrow-derived cells promotes collagenaccumulation in mouse atherosclerotic plaques. Circulation 2008;117:931–9.

[45] Vihinen P, Ala-aho R, Kahari VM. Matrix metalloproteinases as therapeutic targetsin cancer. Curr Cancer Drug Targets 2005;5:203–20.

[46] Newby AC. Dual role of matrix metalloproteinases (matrixins) in intimalthickening and atherosclerotic plaque rupture. Physiol Rev 2005;85:1–31.

[47] Fedak PW, Altamentova SM, Weisel RD, Nili N, Ohno N, Verma S, et al. Matrixremodeling in experimental and human heart failure: a possible regulatory rolefor TIMP-3. Am J Physiol Heart Circ Physiol 2003;284:H626–34.

[48] Wang J, Medina C, Radomski MW, Gilmer JF. N-substituted homopiperazinebarbiturates as gelatinase inhibitors. Bioorg Med Chem 2011;19:4985–99.


http://refhub.elsevier.com/S0969-8051(14)00197-8/rf0005

















http://dx.doi.org/10.1016/j.ijcard.2014.04.071










http://dx.doi.org/10.1038/nm.3437














































































http://dx.doi.org/10.1016/j.ijcard.2014.04.071



































Documents

Gene expression levels of matrix metalloproteinases in human atherosclerotic plaques and evaluation of radiolabeled inhibitors as imaging agents for plaque vulnerability