European Journal of Nuclear Medicine and Molecular Imaging

, Volume 34, Issue 6, pp 830–840

Molecular imaging of vascular cell adhesion molecule-1 expression in experimental atherosclerotic plaques with radiolabelled B2702-p


  • A. Broisat
    • INSERM, U340, Radiopharmaceutiques Biocliniques
    • Université de Grenoble
    • INSERM, U340, Radiopharmaceutiques Biocliniques
    • Université de Grenoble
  • V. Ardisson
    • INSERM, U340, Radiopharmaceutiques Biocliniques
    • Université de Grenoble
  • D. Boturyn
    • Université de Grenoble
    • LEDSS V - Ingénierie Moléculaire, CNRS UMR 5616
  • P. Dumy
    • Université de Grenoble
    • LEDSS V - Ingénierie Moléculaire, CNRS UMR 5616
  • D. Fagret
    • INSERM, U340, Radiopharmaceutiques Biocliniques
    • Université de Grenoble
  • C. Ghezzi
    • INSERM, U340, Radiopharmaceutiques Biocliniques
    • Université de Grenoble
Molecular Imaging

DOI: 10.1007/s00259-006-0310-4

Cite this article as:
Broisat, A., Riou, L.M., Ardisson, V. et al. Eur J Nucl Med Mol Imaging (2007) 34: 830. doi:10.1007/s00259-006-0310-4



VCAM-1 plays a major role in the chronic inflammatory processes present in vulnerable atherosclerotic plaques. The residues 75–84 (B2702-p) and 84–75/75–84 (B2702-rp) of the major histocompatibility complex-1 (MHC-1) molecule B2702 were previously shown to bind specifically to VCAM-1. We hypothesised that radiolabelled B2702-p and B2702-rp might have potential for the molecular imaging of vascular cell adhesion molecule-1 (VCAM-1) expression in atherosclerotic plaques.


Preliminary biodistribution studies indicated that 125I-B2702-rp was unsuitable for in vivo imaging owing to extremely high lung uptake. 123I- or 99mTc-labelled B2702-p was injected intravenously to Watanabe heritable hyperlipidaemic rabbits (WHHL, n = 6) and control animals (n = 6). After 180 min, aortas were harvested for ex vivo autoradiographic imaging, gamma-well counting, VCAM-1 immunohistology and Sudan IV lipid staining.


Robust VCAM-1 immunostaining was observed in Sudan IV-positive and to a lesser extent in Sudan IV-negative areas of WHHL animals, whereas no expression was detected in control animals. Significant 2.9-fold and 1.9-fold increases in 123I-B2702-p and 99mTc-B2702-p aortic-to-blood ratios, respectively, were observed between WHHL and control animals (p < 0.05). Tracer uptake on ex vivo images co-localised with atherosclerotic plaques. Image quantification indicated a graded increase in 123I-B2702-p and 99mTc-B2702-p activities from control to Sudan IV-negative and to Sudan IV-positive areas, consistent with the observed pattern of VCAM-1 expression. Sudan IV-positive to control area tracer activity ratios were 17.0 ± 9.0 and 5.9 ± 1.8 for 123I-B2702-p and 99mTc-B2702-p, respectively.


Radiolabelled B2702-p is a potentially useful radiotracer for the molecular imaging of VCAM-1 in atherosclerosis.


AtherosclerosisNuclear imagingPlaqueVCAM-1


Great advances have been made over the past decade in the understanding of the mechanisms underlying atherogenesis [15], and rupture or erosion of a vulnerable coronary atherosclerotic plaque and subsequent thrombus formation are now recognised as the main events leading to a major cardiovascular complication [6]. However, at present no non-invasive diagnostic and prognostic tool is available for routine use in the detection of prone-to-rupture coronary plaques [7].

Vulnerable coronary plaques are characterised by a thin fibrous cap surrounding a large lipidic and necrotic core and by intense ongoing inflammation. The inflammatory process leading to the development of atherosclerotic lesions towards a vulnerable phenotype is characterised by extensive recruitment of monocytes and lymphocytes into the arterial wall [3]. Several endothelial adhesion molecules are implicated in the process of leucocyte rolling, firm adhesion and transmigration, such as E- and P-selectins, vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1) [8, 9]. VCAM-1 is a 110-kDa molecule of the immunoglobulin family which binds to the very late antigen-4 (VLA-4) integrin present on the surface of leucocytes [10] and the expression of which by endothelial cells is initially triggered by oxidised low-density lipoproteins under atherogenic conditions [1113]. Previous studies have demonstrated the critical role of VCAM-1 in the initiation of atherosclerotic lesions and their progression towards vulnerable plaques [1416]. Moreover, VCAM-1 is strongly expressed by vulnerable plaques and this overexpression is restricted to atherosclerotic lesions whereas the expression of the adhesion molecule ICAM-1 extends into uninvolved areas [14]. As active inflammation characterised by monocytes and T cell infiltration is recognised as a major criterion for defining a vulnerable plaque [6], the adhesion molecule VCAM-1, which is required for the recruitment of these inflammatory cells, is a potential molecular target for the non-invasive detection of vulnerable plaques.

Ling et al. [17] demonstrated that the residues 75–84 and the repeated sequence 84–75/75–84 pertaining to the MHC-I molecule B2702 specifically bound to VCAM-1. Specific binding of the B2702(75–84) peptide (B2702-p) and B2702(84–75/75–84) repeated peptide (B2702-rp) to VCAM-1 was found to be the mechanism by which these peptides had profound immunomodulatory effects both in vitro and in vivo in animal and human models [18]. Considering that VCAM-1 is overexpressed in vulnerable plaques, the hypothesis tested in the present study was that B2702-p and B2702-rp would bind to VCAM-1-expressing atherosclerotic plaques in vivo and that the radiolabelling of these sequences would allow plaque localisation using molecular nuclear imaging.

Materials and methods

In vitro experiments: fluorescence polarisation

Binding of B2702-p to VCAM-1 was evaluated in vitro by fluorescence polarisation using a fluorescent analog of B2702-p ([F]-B2702-p). The fluorescence polarisation of a 12.5 nmol/l (100 μl) solution of [F]-B2702-p was measured in the absence of VCAM-1 or following incremental additions (0.5 μl) of a 13.5 μmol/l solution of the adhesion molecule using a Perkin-Elmer LS 50 spectrometer. Control experiments were conducted by replacing VCAM-1 by bovine serum albumin as a non-specific ligand (BSA, final concentration ranging from 0 to 27 μmol/l). The experiments were performed in triplicate. Anisotropy data (A, arbitrary units) were fitted using the formula \(A = A_{0} + {\left( {A_{{\max }} - A_{0} } \right)} \times {\left( {{{\left[ {{\text{Target}}} \right]}} \mathord{\left/ {\vphantom {{{\left[ {{\text{Target}}} \right]}} {{\left( {K_{{\text{d}}} + {\left[ {{\text{Target}}} \right]}} \right)}}}} \right. \kern-\nulldelimiterspace} {{\left( {K_{{\text{d}}} + {\left[ {{\text{Target}}} \right]}} \right)}}} \right)} \) for a one-to-one interaction, where A0 represents the anisotropy value in the absence of VCAM-1 and Amax, the maximum anisotropy value that was observed in the presence of increasing concentrations of the molecular target (VCAM-1 or BSA) [19]. The Kd value for the interaction between [F]-B2702-p and VCAM-1 or BSA was determined according to the above equation.

In vivo experiments: animal model

Watanabe heritable hyperlipidaemic (WHHL rabbits; WHHL group, n = 16) and control (New Zealand) rabbits (control group, n = 15) were obtained from the Centre de Production Animale (Olivet, France). WHHL and control rabbits were 18.3 ± 2.3 months and 17.5 ± 0.8 months old, respectively (p = NS).

Experimental protocols

Preliminary studies

Preliminary studies consisted in a 30-min protocol aimed at determining whether the biodistributions of 125I-B2702-p and 125I-B2702-rp in WHHL (125I-B2702-p, n = 4; 125I-B2702-rp, n = 6) and control rabbits (125I-B2702-p, n = 4; 125I-B2702-rp, n = 5) were compatible with further detailed evaluation of the tracers at 180 min post injection.

B2702-p evaluation

The 180-min biodistributions and aortic uptake of B2702-p radiolabelled with the clinically used isotopes 123I (123I-B2702-p) and 99mTc (99mTc-B2702-p) were then determined in WHHL (123I-B2702-p, n = 3; 99mTc-B2702-p, n = 3) and control rabbits (123I-B2702-p, n = 3; 99mTc-B2702-p, n = 3).

In both protocols, the animals were anaesthetised with an intramuscular injection of xylazine (5 mg/kg) and ketamine (35 mg/kg). Additional intramuscular injections were performed as necessary throughout the protocol. Two catheters (Introcan-W, Braun, Boulogne, France) were inserted into the ear veins for fluid and tracer administration and blood sample collection. Before the initiation of the protocol, 1 ml of blood was collected from the ear vein and centrifuged for 2 min at 4,000 rpm, and the plasma was frozen for subsequent enzymatic determination of cholesterol and triglyceride concentrations at the clinical laboratory of the University Hospital of Grenoble, Grenoble, France.

The radiotracers were then injected through the ear vein. The injected doses (ID) were 4.7 MBq/kg and 18.5 MBq/kg for 125I-labelled and 123I- or 99mTc-labelled compounds, respectively.

Blood samples were collected from the contralateral ear vein at 1, 2, 5, 10, 15, 20, 25 and 30 (preliminary studies) or 1, 2, 5, 10, 15, 20, 25, 30, 60, 120 and 180 min following tracer injection in order to determine the blood kinetics of each tracer. The last blood sample was also used for the evaluation of radiotracer stability using high-performance liquid chromatography (HPLC) with acetonitrile/H2O as the eluant following centrifugation of whole blood in Nanosep 10 K Omega (Pall life Science, New York, NY). Thirty (preliminary studies) or 180 min after tracer injection, the animals were euthanised with an overdose of a 1:1 solution of pentobarbital and KCl (10% wt/vol) and samples of blood, heart, lung, liver, distal thoracic aorta, spleen, kidney, skeletal muscle, stomach and abdominal fat were obtained and rapidly washed and weighed.

Post-mortem analysis

Biodistribution and blood kinetics

Tissue and blood tracer activities were assessed using a gamma-well counter (Cobra II, Packard Instruments, Rungis, France). The window settings were as follows: 125I, 15–75 keV; 123I, 85–200 keV; 99mTc, 100–168 keV. Blood and tissue activities were expressed as a percentage of the injected dose per gram (%ID/g).


Samples of thoracic and abdominal aortas were embedded in paraffin for immunohistochemical analysis. Serial 5-μm-thick sections were obtained and stained with a standard trichrome, haematoxylin-eosin-safran (HES) method or immunostained using an anti-VCAM-1 antibody. Tissue slices adjacent to those selected after trichrome colouration were used for immunostaining. After endogenous avidin and biotin blockade, sections were incubated at 4°C overnight with the primary anti-VCAM-1 polyclonal goat antibody (1:500; Santa Cruz Biotechnology). Incubations for 1 h at room temperature with the secondary biotinylated anti-goat IgG horse antibody (1:400, Vector Laboratories), and for 30 min with avidin peroxidase were then performed. Peroxidase was visualised with diaminobenzidine as the chromogen and sections were then counterstained with haematoxylin. Control slides were stained in the absence of the primary, anti-VCAM-1 antibody and were free of peroxidase activity, indicating the specificity of VCAM-1 immunostaining. Digital pictures were obtained at a 100 ×  magnification under a Nikon Diaphot microscope.


At the end of the 180-min protocol, the thoracic and abdominal aortas were quickly removed and placed in 10% buffered formalin (Accustain, Sigma-Aldrich) for the time necessary for the removal of extravascular tissue (∼5–8 min). The adventitial tissue was thoroughly removed and the aortas were opened longitudinally. The arteries were covered with plastic wrap and placed on a phospho-imager film (Fujifilm 08SR2025) for 18 h before scanning (Fujifilm BAS 5000). Quantification of the autoradiographic images was performed using dedicated software (Image-Gauge, Fuji). Based on the results from Sudan IV staining of lipid-rich areas (see below), three regions of interest (ROIs; ∼0.5 × 0.5 cm each) were drawn within the Sudan IV-negative and Sudan IV-positive areas of WHHL animals, corresponding respectively to areas without and with atherosclerotic plaques (see below). In control animals, six ROIs were defined along the aorta. Tracer activities from image quantification were expressed as phospho-stimulated luminescence units/mm2 (PSL/mm2) and were normalised to the ID expressed in MBq/kg (PSL/mm2/ID). Note that the phospho-imager film exposure and therefore the PSL values are dependent upon the radionuclide used, therefore precluding direct comparisons between 123I and 99mTc activities from quantification of autoradiographic images.

Sudan IV colouration

Sudan IV lipid staining was performed as previously described on a similar experimental model [20]. Following autoradiographic image acquisition, the aortas were immersed for 10 min in a Sudan IV solution (200 mg of Sudan IV in 100 ml of 1/1 70% ethanol/acetone) and for 5 min in 50% ethanol in order to bleach the unstained areas not containing lipids. Atherosclerotic plaques containing lipids were therefore stained red (Sudan IV-positive areas) whereas normal regions of the aortas were left unstained (Sudan IV-negative areas). Digital pictures (Nikon D70) were obtained following Sudan IV staining and used for the delineation of Sudan IV-positive areas corresponding to areas of atherosclerotic plaque development using Photoshop CS software. The outline of the vessel and Sudan IV-positive areas was then superimposed on the autoradiographic image to ensure that tracer uptake preferentially occurred within Sudan IV-positive areas. ROIs corresponding to Sudan IV-positive and -negative areas were drawn within and outside of these areas, respectively.

Chemistry and radiochemistry

Peptide synthesis

B2702-p [RENLRIALRY], B2702-rp [YRLAIRLNERRENLRIALRY] and a modified B2702-p for technetium labelling [HGRENLRIALRY] were prepared by solid-phase synthesis. Assembly of all protected peptides was performed either manually using a Fmoc/tBu strategy in a glass reaction vessel fitted with a sintered glass frit or automatically on a synthesiser (348 Ω synthesiser, Advance ChemTech). Coupling reactions were performed using 1.5–2 eq. of N-α-Fmoc protected amino acid activated in situ with 1.5–2 eq. PyBOP and 3–4 eq. DIEA in DMF (10 ml/g resin) for 30 min (proportions are given relative to resin loading). When performed manually, syntheses were controlled by Kaiser and/or TNBS tests. N-α-Fmoc protecting groups were removed by treatment with a piperidine:DMF solution (1:4) (10 ml/g resin) for 10 min. The process was repeated three times and the completeness of unprotection was checked by the UV absorption of the piperidine washings at 299 nm.

Synthetic linear peptides were recovered directly upon acid cleavage (1% TFA in CH2Cl2). The resins were treated for 3 min repeatedly until the resin beads turned dark purple. The combined washings were concentrated under reduced pressure and white solid peptides were obtained by precipitation from ether and analysed by reversed-phase HPLC. When necessary, purification was performed on a preparative column.


125I labelling of B2702-rp and 125I or 123I labelling of B2702-p were performed on the tyrosine residue (#84) using chloramine-T. Briefly, 74 MBq of 125I or 200 MBq of 123I was added to 20 or 40 nmol of peptide and 20 or 80 μl of extemporaneously prepared chloramine T (1 mg/ml), respectively. 125I and 123I were purchased from Amersham Radiochemical Centre and Schering SA, respectively. The reaction was stopped after 15 min by the addition of 20 μl or 80 μl of sodium metabisulphide (4 mg/ml) for 125I and 123I, respectively. Thin layer chromatography (TLC) analysis (RP-18R254; Merck) with acetonitrile/H2O (60/40) as the eluent indicated a radiochemical purity (RCP) of >95% for 125I and >80% for 123I. When 123I-B2702-p RCP was between 80% and 90%, the solution containing the tracer was purified using an anion exchange column with a phosphate buffer solution (pH = 7.4) as the eluent.

Two amino acids [HG] were added to B2702-p for 99mTc labelling on the histidine with [99mTc(OH2)3 (CO)3] using a tridentate ligand system. The precursor [99mTc(OH2)3 (CO)3] was synthesised using a tricarbonyl pharmaceuticals Kit (Isolink, Mallinckrodt). The kit was reconstituted with 2 GBq of 99mTcO4 (Schering SA), and incubated at 100°C for 20 min. After pH adjustment to 8.0, 800 MBq of this solution was added to 30 nmol of B2702 and incubated for 20 min at 80°C. Radiochemical yield was assessed by TLC (RP-18R254; Merck) with either acetonitrile/H2O (60/40) or NaCl 0.9% as the solvent. RCP was >90%. Finally, fluorescent labelling of B2702-p was performed by addinga fluorescein molecule at the N-terminal end of the peptide.

Statistical analysis

Results were expressed as mean±SEM. Statistical comparisons were performed using a Mann and Whitney test except for the blood kinetics of the tracers in control and WHHL animals, which were compared using an ANOVA test. A p value<0.05 was considered statistically significant.


Fluorescence polarisation

Results from fluorescence polarisation experiments are presented in Fig. 1. An increasing anisotropy value was observed in the presence of increasing VCAM-1 concentrations and a plateau was reached. A Kd value of 2.7 × 10−7 mol/l for the interaction between VCAM-1 and [F]-B2702 was determined assuming a one-to-one interaction (see Materials and methods). This value was ~600-fold lower than that observed in the presence of BSA (1.6·10−4 mol/l).
Fig. 1

Anisotropy values from [F]-B2702-p in the presence of increasing concentrations of VCAM-1. Results from control experiments performed in the absence of VCAM-1 and in the presence of the non-specific target BSA are shown in the insert

Experimental model validation

Plasma measurements

Serum cholesterol and triglyceride concentrations were 0.42 ± 0.03 g/l and 0.35 ± 0.03 g/l in control animals and 6.62 ± 0.32 g/l and 2.70 ± 0.39 g/l in WHHL animals, respectively (p < 0.001 vs control).

Sudan IV lipid staining

Sudan IV stained lipid-containing areas in red whereas areas with no lipid accumulation were left unstained. No atherosclerotic lesions were detected on aortas from control animals following lipid staining with Sudan IV whereas red, lipid-containing atherosclerotic lesions were readily observed on aortas from WHHL animals. The most extensive lesions were observed on the ascending and proximal part of the thoracic aorta whereas lesions on the distal thoracic and abdominal aorta were mainly observed at branching sites.

Histochemistry and immunohistochemistry

Figure 2 presents the standard, trichrome colouration and VCAM-1 immunohistochemistry from aortas of control animals (a and d, respectively) and of WHHL animals in Sudan IV-negative (b and e, respectively) and Sudan IV-positive areas (c and f, respectively). Advanced atherosclerotic lesions containing a large lipid core were found in the thoracic and abdominal aortas of WHHL animals following standard trichrome colouration. VCAM-1 protein expression as revealed by the specific brown diaminobenzidine staining was observed on the endothelium overlying atherosclerotic lesions of the thoracic and abdominal aortas as well as in the subendothelial intimal space and in the necrotic core of advanced lesions of WHHL animals (Sudan-positive areas, Fig. 2f). VCAM-1 expression was also observed on the endothelium of Sudan IV-negative areas on aortas from WHHL rabbits (Fig. 2e), whereas no VCAM-1 expression was detected in aortas from control animals (Fig. 2d).
Fig. 2

Histochemistry and immunohistochemistry. a–c Standard HES staining. No atherosclerotic lesions were observed in aortas from the control group (a) whereas early (b) and advanced lesions (c) were found in the WHHL thoracic and abdominal aortas following HES staining. df Immunohistochemistry. No VCAM-1 expression was detected in aortas from control animals (d). VCAM-1 expression was observed on the endothelium of Sudan IV-negative areas on aortas from WHHL rabbits (e, arrow). More robust VCAM-1 immunostaining was observed on the endothelium overlying atherosclerotic lesions (f, arrow) of the thoracic and abdominal aortas as well as in the subendothelial intimal space and the necrotic core of advanced lesions of WHHL animals (f, arrowheads). gi Specificity of the immunostaining was assessed by performing control experiments in the absence of the primary antibody

Preliminary studies

Organ biodistributions of 125I-B2702-p and 125I-B2702-rp 30 min following tracer injection in control and WHHL animals are presented in Table 1. The mean aortic activity of 125I-B2702-p and 125I-B2702-rp was increased by 2.0- to 2.5-fold in WHHL rabbits when compared with control rabbits (p < 0.05 vs control for 125I-B2702-p). No significant difference was observed when comparing 125I-B2702-p or 125I-B2702-rp activity between control and WHHL animals in the other organs evaluated. Table 1 also indicates an extremely high 125I-B2702-rp lung activity in both the control and the WHHL group (7.1 ± 2.8%ID/g and 9.1 ± 2.1%ID/g, respectively, p = NS) as compared with 125I-B2702-p (0.3 ± 0.01%ID/g and 0.2 ± 0.0%ID/g, respectively). 125I-B2702-rp was mainly excreted through the hepatobiliary route, whereas 125I-B2702-p excretion occurred mainly through the kidneys.
Table 1

Biodistribution of 125I-B2702-rp and 125I-B2702-p 30 min following tracer injection




Control (n = 5)

WHHL (n = 6)

Control (n = 4)

WHHL (n = 4)


0.083 ± 0.012

0.079 ± 0.013

0.270 ± 0.006

0.303 ± 0.029


0.043 ± 0.011

0.045 ± 0.007

0.108 ± 0.007

0.120 ± 0.011


0.010 ± 0.003

0.026 ± 0.005

0.027 ± 0.006

0.057 ± 0.016*


7.095 ± 2.765

9.104 ± 2.127

0.343 ± 0.096

0.217 ± 0.019


0.414 ± 0.147

0.331 ± 0.053

0.148 ± 0.009

0.174 ± 0.013


0.783 ± 0.352

0.476 ± 0.180

0.100 ± 0.006

0.125 ± 0.011


0.355 ± 0.145

0.250 ± 0.054

0.391 ± 0.065

0.518 ± 0.064


0.007 ± 0.002

0.009 ± 0.001

0.029 ± 0.006

0.018 ± 0.003


0.014 ± 0.003

0.010 ± 0.002

0.044 ± 0.003

0.041 ± 0.002


0.049 ± 0.007

0.038 ± 0.007

0.119 ± 0.009

0.121 ± 0.017

Uptake values are expressed as %ID/g

*p < 0.05 vs control

123I-B2702-p and 99mTc-B2702-p evaluation

Biodistribution and blood kinetics

At 180 min, there was a non-significant 3.1-fold and 1.8-fold increase in 123I-B2702-p and 99mTc-B2702-p aortic activity, respectively, from control to WHHL animals (Table 2). As observed at the 30-min time point, mean aortic-to-blood ratios were significantly higher in WHHL than in control animal (Fig. 3). No significant difference was observed in 123I-B2702-p or 99mTc-B2702-p activity between control and WHHL animals in all organs evaluated, and the kidneys were the major organ for 123I-B2702-p and 99mTc-B2702-p excretion (Table 2).
Table 2

Biodistribution of 123I-B2702-p and 99mTc-B2702-p 180 min following tracer injection





Control (n = 3)

WHHL (n = 3)

Control (n = 3)

WHHL (n = 3)


0.190 ± 0.038

0.161 ± 0.048

0.097 ± 0.017

0.102 ± 0.033


0.088 ± 0.017

0.088 ± 0.031

0.029 ± 0.008

0.044 ± 0.015


0.011 ± 0.004

0.034 ± 0.012

0.025 ± 0.009

0.046 ± 0.012


0.138 ± 0.028

0.163 ± 0.044

0.166 ± 0.052

0.208 ± 0.044


0.086 ± 0.025

0.096 ± 0.029

0.427 ± 0.080

0.332 ± 0.042


0.076 ± 0.015

0.089 ± 0.029

0.198 ± 0.061

0.172 ± 0.070


0.307 ± 0.052

0.490 ± 0.184

0.558 ± 0.109

1.361 ± 0.721


0.011 ± 0.001

0.015 ± 0.006

0.012 ± 0.001

0.019 ± 0.009


0.028 ± 0.003

0.026 ± 0.009

0.008 ± 0.002

0.017 ± 0.003


0.095 ± 0.025

0.114 ± 0.040

0.045 ± 0.009

0.052 ± 0.015

Uptake values are expressed as %ID/g
Fig. 3

Mean aortic-to-blood tracer activity ratios from gamma-well counting 180 min following injection of 123I-B2702-p and 99mTc-B2702-p. *p < 0.05 vs control

The 180-min blood kinetics of 123I-B2702-p and 99mTc-B2702-p are presented in Fig. 4. For each tracer, no significant difference was observed between the control and WHHL group following ANOVA analysis. Blood clearances were biexponential, with an initial fast component (T1/2 = 2–4 min for both tracers) followed by a slower component (T1/2∼120 min for 99mTc-B2702-p and ∼600 min for 123I-B2702-p). At 180 min after injection, circulating 123I-B2702-p and 99mTc-B2702-p activity represented 32.9 ± 3.5% and 12.8 ± 1.8% of the initial, 1-min activity (p < 0.05 for both comparisons), respectively.
Fig. 4

Blood kinetics of 123I-B2702-p and 99mTc-B2702-p over 180 min

Ex vivo imaging of VCAM-1 expression

Depicted on Fig. 5 are representative examples of Sudan IV colouration and ex vivo autoradiographic images of 99mTc-B2702-p 180 min after tracer injection to a control (a and b) and a WHHL rabbit (c and d). The aorta from the control animal did not exhibit lipid staining (a) and 99mTc-B2702-p activity on the ex vivo image was low and homogeneous (b). Atherosclerotic lesions were observed on the aorta from the WHHL animal (c), and tracer uptake on ex vivo images co-localised with Sudan IV-positive, lipid-rich areas corresponding to atherosclerotic plaques (d). Image quantification is shown in Fig. 6. At 180 min after injection, 123I-B2702-p and 99mTc-B2702-p activities on Sudan-positive areas of the WHHL group were 2.33 ± 1.23 PSL/mm2 and 0.25 ± 0.08 PSL/mm2, respectively. These activities were 17.0- and 5.9-fold greater than those observed in the control group (0.14 ± 0.01 and 0.04 ± 0.01, respectively, p < 0.05 for both comparisons). Tracer activities were also significantly higher in Sudan-positive vs Sudan-negative areas within the same animals, with 123I-B2702-p and 99mTc-B2702-p activity ratios of 4.0 and 4.8, respectively. Finally, there was a 4.3-fold difference between 123I-B2702-p activity in Sudan-negative areas of WHHL animals as compared with that observed in control animals (p < 0.05).
Fig. 5

Representative examples of Sudan IV lipid staining and ex vivo autoradiographic imaging from animals injected with 123I-B2702-p. a, b Representative example of Sudan IV lipid staining of the thoracic aorta from a control animal (a) and the corresponding autoradiographic image (b). Sudan IV staining was negative and tracer uptake was low and homogeneous. c, d Representative example of Sudan IV lipid staining of the thoracic aorta from a WHHL animal (c) and the corresponding autoradiographic image (d). Preferential tracer uptake was observed at sites of lipid accumulation corresponding to advanced atherosclerotic lesions. The delineation of Sudan IV-stained areas from the vessel shown in c is depicted in e. This outlining was then superimposed on the autoradiographic image (f) to confirm the co-registration between areas of plaque development and sites of preferential tracer uptake, and image quantification was performed as described in Materials and methods
Fig. 6

Quantification of 123I-B2702-p (a) and 99mTc-B2702-p (b) aortic activities from ex vivo autoradiographic imaging 180 min following tracer injection in control animals and in Sudan IV-negative and Sudan IV-positive areas of WHHL animals. *p < 0.05 vs control,p < 0.05 vs Sudan IV-negative area

Stability of tracers

Following incubation of the tracers for 60 min in whole blood and centrifugation in Nanosep 10 K Omega, 123I-B2702 and 99mTc-B2702 activities were predominantly found in the plasma (63% and 85% whole blood activity, respectively). Unbound 123I-B2702 and 99mTc-B2702 plasma activity represented 82% and 85% of the total plasma activity, respectively. Moreover, HPLC analysis indicated that free 123I and free 99mTc represented 20–30% of the total blood radioactivity at 180 min following injection of 123I-B2702-p and 99mTc-B2702-p, respectively, and that the remaining 70–80% of the activity exclusively consisted of 123I-B2702-p or 99mTc-B2702-p.


Vulnerable atherosclerotic plaques are characterised by a thin fibrous cap surrounding a large lipidic and necrotic core and by an intense and chronic inflammatory process. Plaque rupture or erosion and subsequent thrombus formation are the leading cause of coronary events [3, 6]. Over the course of the inflammatory process, cell-to-cell and cell-to-matrix interactions are mediated by a large range of adhesion molecules which allow monocyte and T cell recruitment into the arterial wall [8, 9]. Among these adhesion molecules, VCAM-1 plays a major role in the initiation and the development of atherosclerotic lesions [15, 16]. This 110-kDa immunoglobulin-like protein is expressed on endothelial cells subjected to pro-atherogenic conditions. VCAM-1 binds to the VLA-4 integrin, which is constitutively expressed on lymphocytes and monocytes [10]. Cybulsky et al. [16] generated homozygous VCAM-1 domain 4-deficient mice (VCAM-1D4D/D4D) and observed that atherosclerotic plaque area was reduced by 40% in VCAM-1D4D/D4D low-density lipoprotein receptor (LDLR)−/− mice aorta when compared with atherosclerotic LDLR−/− mice, indicating that VCAM-1 is critical in atherogenesis. Moreover, VCAM-1 is overexpressed on vulnerable atherosclerotic plaques and this expression is restricted to plaques and plaque-predisposed regions in rabbits [12, 14], mice [14] and humans [21, 22]. VCAM-1 expression is strongly correlated to macrophage and lymphocyte accumulation [21, 22], which represent well-accepted features of a vulnerable plaque [6]. VCAM-1 is therefore a potential target for the non-invasive molecular imaging of vulnerable plaques.

In the present study, the VCAM-1 ligands B2702-p and B2702-rp were radiolabelled and their ability to image atherosclerotic plaques was evaluated in a hyperlipidaemic rabbit model of atherosclerosis. B2702-p and B2702-rp belong to a series of peptides corresponding to linear sequences of an MHC-I molecule with profound immunomodulatory effects in vitro and in vivo in both animal and human models [18]. One of these peptides corresponding to the α1 helix of MHC-I B2702 was shown to block cell-mediated cytotoxicity. The observed immunomodulatory effects were attributed to the carboxyl terminal 10 amino acids 75–84 peptide (B2702-p), and the repeated sequence 84–75/75–84 (B2702-rp) was the most efficient. In mice, an ovalbumin-induced allergic pulmonary response was blocked by in vivo administration of either B2702-rp or an anti-VLA-4 antibody [17]. The authors provided direct evidence that the inhibition of leucocyte recruitment from peripheral blood by B2702-rp was due to direct binding of B2702-rp to VCAM-1.

The experiments presented in this study were performed on Watanabe heritable hyperlipidaemic (WHHL) rabbits, a model of familial hypercholesterolaemia [23, 24] in which spontaneous development of atherosclerotic plaques containing large amounts of leucocytes, monocytes and foam cells was previously documented [25].

In the present study, VCAM-1 expression was assessed using immunohistochemistry in WHHL animals and was observed on the endothelium as well as in the deeper layer of the intima and within the necrotic core of advanced atherosclerotic lesions identified using Sudan IV staining. VCAM-1 expression in WHHL rabbits was also observed on the endothelium of regions not overlying a well-developed lesion (Sudan IV-negative areas) and corresponding to an early stage of endothelial activation. This pattern of VCAM-1 expression is in accordance with previously published results from Fruebis et al. [26] and Iiyama et al. [14] on a similar experimental model. The authors demonstrated that VCAM-1 found within atherosclerotic lesions was predominantly expressed by endothelial cells in early lesions but also by intimal proliferating smooth muscle cells and to a smaller extent by the endothelium from the neovasculature and by macrophages in more advanced lesions. The results from the present study also indicated the absence of detectable expression of VCAM-1 in control, normolipidaemic animals.

The affinity of B2702-p for VCAM-1 was determined in vitro using fluorescence polarisation and a fluorescent analogue of B2702-p ([F]-B2702-p) as previously described [19]. The results indicated a Kd value of 2.7 × 10−7 mol/l for the interaction between the fluorescent probe and VCAM-1. Preliminary in vivo biodistribution studies indicated that iodine-labelled B2702-rp was predominantly localised in the lungs of WHHL and control animals 30 min after tracer injection. The higher lung uptake of 125I-B2702-rp as compared with that of 125I-B2702-p precludes its potential use for in vivo imaging of coronary atherosclerosis and was likely due to the lipophilic properties of the dimeric peptide. Biodistribution studies also indicated that, in accordance with the lipophilic nature of the tracer, 125I-B2702-rp was mainly excreted through the hepatobiliary route whereas the more hydrophilic tracer, 125I-B2702-p, was predominantly eliminated by the kidneys. Given the considerable 125I-B2702-rp lung activity, B2702-p was selected for further detailed evaluation over 180 min following radiolabelling with 123I or 99mTc. The fact that the trend towards a higher 123I-B2702-p or 99mTc-B2702-p aortic activity in WHHL animals as compared with control rabbits was not statistically significant was probably due to the evaluation of global aortic tracer activity, which represented the mean activity present in both Sudan IV-positive and Sudan IV-negative areas. It is therefore likely that the most intense tracer uptake was diluted over a larger tissue mass. Moreover, the size of atherosclerotic plaques in WHHL rabbits decreases from the ascending to the abdominal aorta [24], and aortic samples dedicated to gamma-well counting were obtained from the distal part of the thoracic aorta in order to perform autoradiographic imaging on more proximal, severely injured aortic areas. However, the uptake of 123I-labelled or 99mTc-labelled B2702-p as measured from biodistribution studies and expressed as %ID/g are comparable to those recently described using 99mTc-annexin A5 in a similar experimental model [27]. Normalisation of the tracers’ aortic uptake to blood activity resulted in significantly higher 123I-B2702-p and 99mTc-B2702-p aortic-to-blood ratios in WHHL than in control animals (0.45 ± 0.03 and 0.24 ± 0.06, respectively, for 99mTc-B2702-p), indicating preferential binding of the tracers to VCAM-1-expressing vessels. However, considering the great importance of the lesion-to-blood ratio in the detection of small atherosclerotic lesions [28], further studies will have to address whether true radiolabelled B2702-p lesion-to-blood activity ratios compare favourably with those already described for 99mTc-annexin A5 (3.0 ± 0.37) [29].

These results were confirmed by ex vivo autoradiographic imaging, which indicated that 123I-B2702-p and 99mTc-B2702-p accumulated at sites of atherosclerotic lesions identified using Sudan IV staining of the aortas from WHHL animals. Tracer activities in Sudan IV-positive areas were significantly higher than those observed in remote regions of the aortas of WHHL animals and in aortas from control, normolipidaemic rabbits. Moreover, a graded increased in 123I-B2702-p activity from the aortas of control animals to the Sudan-negative and Sudan-positive areas of the aortas of WHHL animals was also observed, in accordance with the VCAM-1 expression pattern described above. Although the possibility exists that the immersion of the vessels in 10% buffered formalin prior to autoradiographic imaging, as previously described [30], caused a slight washout of the radioactivity from the tissue, the presence of the fixative formalin may have strengthened the interaction between radiolabelled B2702-p and VCAM-1, thereby preventing the loss of tracer following vessel excision. In any case, all animals were treated similarly and this step should therefore not have interfered with the comparisons between control and WHHL animals. Moreover, direct evidence for the co-localisation of the tracer and VCAM-1 expression could not be obtained in the present study for technical reasons. Indeed, tissue preparation for immunohistochemical studies, i.e. tissue fixation and paraffin embedding, must be started immediately following tissue excision and last for a period of time not compatible with subsequent microautoradiographic imaging. In addition, multiple incubations of tissue samples in dehydrating solutions might have potentially affected the amount of originally bound tracer. The quantification of ex vivo images indicated that the 123I-B2702-p plaque-to-control ratio was threefold higher than the corresponding 99mTc-B2702-p ratio 180 min after injection. This difference may have been partly due to the fact that 123I-B2702-p is radiolabelled at its C-terminal end whereas 99mTc-B2702-p is modified and radiolabelled at its N-terminal side, suggesting that the specificity of B2702-p binding to VCAM-1 may have been affected by modifications performed at the N-terminal end of the peptide. Although the in vitro affinity of 99mTc-B2702-p for VCAM-1 was not determined, it is likely that this value was similar to that observed for [F]-B2702-p, considering that both fluorescein and 99mTc labelling of B2702-p was performed at the N-terminal end of the molecule.

Imaging atherosclerotic plaque is an emerging challenge. Several invasive and non-invasive imaging techniques are currently under evaluation. Intravascular ultrasound imaging (IVUS), optical coherence tomography (OCT), positron-sensitive probe and thermography are the more promising invasive methodologies that are currently being developed [3134]. Despite the high resolution of magnetic resonance imaging and the high sensitivity of nuclear imaging, together with its potential for functional imaging, as yet no non-invasive technique is available routinely for clinical use [7].

Several radiotracers targeted at oxidised LDL, proliferating smooth muscle cells, macrophages, apoptotic cells, matrix metalloproteinases and thrombus have been synthesised and evaluated [30, 3540]. [18F]FDG accumulation in macrophages within the plaque has also been proposed for in vivo PET imaging [34]. However, while radioactivity was detected on the aortas of WHHL rabbits, the tracer might be suboptimal for coronary plaque imaging owing to the high myocardial background radioactivity, leading the authors to hypothesise that an invasive, intravascular positron-sensitive probe would be better suited for coronary [18F]FDG detection. Although it might seem suboptimal for coronary atherosclerosis imaging, the radiolabelled B2702-p aortic-to-myocardial ratio as estimated from the results of the present study (Table 2) probably does not reflect the true coronary-to-myocardial ratio. Indeed, in addition to the fact that the aortic-to-blood ratio underestimated the actual aortic lesion-to-blood ratio, as discussed above, previous studies have indicated that VCAM-1 is mainly expressed by the newly formed vasa vasorum that develops within atherosclerotic lesions rather than by the arterial lumen [22] and that the coronary bed has a much greater vasa vasorum density than other vascular beds [41]. These studies suggest that radiolabelled B2702-p coronary-to-myocardial ratios will exceed aortic-to-myocardial ratios owing to higher VCAM-1 expression. Further studies are necessary to confirm this hypothesis.

125I- or 99mTc-labelled macrophage chemoattractant protein-1 (MCP-1) bound to CCR-2 receptors expressed on monocytes and macrophages within the plaque in a rabbit model of focally induced atherosclerosis [30, 37]. Ex vivo autoradiographic quantification indicated an average ratio of plaque-to-normal vessel of 6 [30], and atherosclerotic lesions were visible 180 min after injection in vivo [38]. 99mTc-labelled annexin A5 is a radiotracer of apoptosis [28, 29, 39, 42, 43]. An ex vivo autoradiographic study indicated an average ratio of plaque-to-normal vessel of 15 ± 7 [29], and the feasibility of in vivo imaging of apoptosis in porcine coronary arteries with 99mTc-labelled annexin A5 has recently been suggested [39]. In the present study, B2702-p plaque-to-control ratios determined 3 h after injection from quantification of ex vivo images were 17 ± 9 and 5.9 ± 1.8 for 123I-B2702-p and 99mTc-B2702-p, respectively. These ex vivo ratios therefore compare favourably with those obtained by Kolodgie et al. [29] and Ohtsuki et al. [30] with 99mTc-labelled annexin A5 and 125I-MCP-1 (15 ± 7 and 6, respectively).


VCAM-1 plays a major role in the chronic inflammatory processes present in human vulnerable atherosclerotic plaques. In the present study, the VCAM-1 ligand B2702-p, radiolabelled using 123I and 99mTc, allowed the ex vivo localisation of atherosclerotic lesions expressing VCAM-1 on aortas from WHHL rabbits. Further studies using a murine model of carotid atherosclerotic lesion development and a dedicated small animal imaging camera are underway to determine the potential of radiolabelled B2702-p to non-invasively image VCAM-1 expression in vivo.


Financial support was provided by the National Institute for Health and Medical Research (INSERM) and the French Ministry of National Education and Research. All the experimental protocols described in the present study were approved by the Animal Care and Use Committee of the Centre de Recherche et Service de Santé des Armées (CRSSA) and the experiments were performed by an authorised individual (A. Broisat, authorisation # 38 04 37).

Copyright information

© Springer-Verlag 2007