Introduction

Atherosclerotic plaque rupturing is considered an important determinant of the outcome of coronary artery disease. Identifying vulnerable plaques that are prone to rupture is, therefore, of great clinical interest. High concentrations of inflammatory cells, mainly macrophages, has been suggested to be one characteristic of the vulnerable plaques [1–3].

The metabolic activity and inflammatory processes of atherosclerotic plaques in humans and animal models have recently been detected using fluorine-18-labelled fluorodeoxyglucose ([18F]FDG) in positron emission tomography (PET) [4–7]. Despite these promising results, other imaging markers are warranted since [18F]FDG has obvious limitations for cardiac imaging, such as high accumulation to myocardium and dependence on metabolic conditions of the patient.

[11C]PK11195 binds with high affinity and selectivity to peripheral benzodiazepine receptor (PBR, also known as translocator protein) [8, 9]. The PBR is widely expressed in the body and, although it has been found in all cell types studied, the expression levels in different cell types vary. PBR is expressed in high amounts in macrophages, the number of PK binding sites in macrophages is estimated to 250,000 per cell [10, 11]. Little information is known about the numbers of PBR in other cell types, but its relative expression is highest in glandular tissues, lung and myocardium [12, 13] and to a lesser extent in other tissues including the endothelial and smooth muscle cells in the blood vessels [14]. Ligands for PBR have been used successfully for the imaging of degenerative brain diseases, based on the low expression of PBR in the normal brain tissue and high expression of PBR in activated glial cells and infiltrated macrophages. Also, in the periphery, PBR are highly expressed in activated macrophages in the inflamed area.

[11C]PK11195 has previously been found to accumulate into the inflamed lung in animal models [15, 16] and in humans [17] due to the high number of activated macrophages in the inflamed area. A microautography study by Jones et al. [16] shows selective high binding of 3H-labelled PK to macrophages in an inflammatory cell mixture, already at a time point of 10 min. According to these studies, the macrophage express the highest number of PK binding sites among the inflammatory cells. Therefore, [11C]PK11195 has promising qualities as an imaging agent also in atherosclerosis-associated inflammation.

In this study, the feasibility of labelled PK11195 in the assessment of the inflammation in the arterial atherosclerotic plaques was studied both in vitro and ex vivo by using [3H]PK11195 and [11C]PK11195 and aortic sections of a LDLR/ApoB48 atherosclerotic mice. By combining autoradiography with histological and the immunohistological analyses, the accumulation of the tracer was compared between the atherosclerotic plaques and non-atherosclerotic arterial sites.

Materials and methods

Animals

Five male mice deficient for both LDL receptor and ApoB48 (LDLR−/−/ApoB100/100, Jackson Laboratory strain #003000) were investigated. LDLR/ApoB48-deficient mice were chosen as a model because these animals show a high prevalence of atherosclerotic plaques [4, 18–20]. Starting at 10 months of age, the mice were kept on a Western-type diet (Teklad Adjusted Calories, consisting of 21% fat and 0.15% cholesterol without sodium cholate) for 3 months. The C57Bl/6N strain male mice (n = 3) fed with a standard diet were used as non-atherosclerotic controls. The age and weight (mean±SD) of the LDLR/ApoB48 mice were 13 ± 1 months and 39 ± 8 g, respectively, and of the C57BL/6N control mice were 9 months and 28 ± 4 g, respectively. All mice were maintained under standard conditions. Food and water were provided ad libitum. The study plan was approved by the Laboratory Animal Care and Use Committee of the University of Turku, Finland.

Radiotracer synthesis

[11C]PK11195 was prepared as described by Debruyne et al. [21] with slight modifications. [11C]Methyl iodide was let to react with 1.0 mg of desmethyl-PK11195 (ABX, Radeberg, Germany) and 3.0 μL of 1.0 M tetrabutylammonium hydroxide in 150 μL of dimethyl sulfoxide for 3 min at 80°C. The crude product was purified on a semi-preparative μ-Bondapak® C-18 column using 65% of acetonitrile in 10 mM phosphoric acid. After evaporation of the high-performance liquid chromatography (HPLC) solvent, the purified product was formulated in propylene glycol/ethanol, 7/3, and 0.1 M phosphate buffer, pH 7.4, and filtered through a 0.22-μm sterile filter. Measurement of concentration and radiochemical purity of [11C]PK11195 batches was performed by HPLC using an analytical μ-Bondapak® C-18 column, a UV-detector set at 223 nm and a radioactivity detector in series. Radiochemical purity was higher than 99.5%.

In vitro binding of [3H]PK11195 in aorta sections

Twenty-micrometer sections of atherosclerotic mouse aorta were prepared [4] and stored at −20°C. Before use, the sections were warmed to room temperature (RT) and pre-incubated for 5 min at RT in 50 mM Tris–HCl buffer (pH 7.4). Immediately thereafter, the sections were incubated for 30 min at RT with 1.2 nM [3H]PK11195 (85 Ci/mmol; Amersham, SA) in 50 mM Tris–HCl buffer. Non-specific binding was estimated in adjacent sections in the presence of 10 μM (unlabelled) PK11195 (ABX, Radeberg, Germany). After incubation, the slides were washed twice for 10 s with ice-cold buffer and rinsed in distilled water to remove buffer salts. The sections were air-dried and apposed to an imaging plate (Fuji Imaging Plate BAS-TR2025, Fuji Photo Film, Japan) for digital autoradiography. The exposure time was 7 days. The imaging plates were scanned with the Fuji Analyzer BAS-5000. The digital autoradiographic images were analysed for count densities (photo-stimulated luminescence unit per area, PSL/mm2) with a computerised image analysis programme (Tina 2.1, Raytest Isotopenmessgeräte, Straubenhardt, Germany) and the count densities for background areas were subtracted from the image data.

Ex vivo biodistribution of [11C]PK11195 in atherosclerotic and control mice

[11C]PK11195 (23 ± 12 MBq) was given intravenously via a tail vein to non-anaesthetised mice, and after 20 min, the animals were killed by CO2. Blood was immediately collected by cardiac puncture into heparinised tubes, weighed and measured for [11C]radioactivity in a calibrated well counter (NaI(Tl) 3 × 3 in., Bicron 3 MW3/3P, Bicron, Newbury, OH, USA). Tissue samples (aorta, lung, liver and heart) were dissected, patted dry, weighed and their radioactivity were measured. Background counts were subtracted and the radioactivity decay was corrected to the time of injection, and the dose that remained in the tail was compensated. The radioactivity that had accumulated in the tissue samples over the 20-min period following [11C]PK11195 injection was expressed as a percentage of the injected dose per gram of tissue (%ID/g).

Uptake of [11C]PK11195 in aorta sections ex vivo

The distribution of [11C]PK11195 to aorta tissue was studied with digital autoradiography. The aorta from the ascending aorta to the level of the diaphragm was dissected and blood was removed with saline rinsing. The aortas were frozen in mounting media for sectioning with cryomicrotome. Sequential longitudinal 20 μm mouse aorta sections were cut at −15°C, thaw-mounted onto microscope slides, air-dried and apposed to an imaging plate (Fuji Imaging Plate BAS-TR2025, Fuji Photo Film, Japan). After an exposure time of 1 h, the imaging plates were scanned with the Fuji Analyzer BAS-5000 and the images were analysed for count densities (PSL/mm2) with an image analysis programme (Tina 2.1, Raytest Isotopenmessgeräte, Straubenhardt, Germany). Three types of regions of interest (ROI) were defined according to the histology [22], and the background area count densities were subtracted from the image data. The ROIs were (1) healthy vessel wall (control site, no visual progression of atherosclerosis), (2) non-inflamed plaque (mostly core areas without cell infiltration or acellular areas in the plaque) and (3) inflamed plaque (cell-rich areas in the plaques containing mostly macrophages). The ROI were drawn with the help of a microscope on carefully co-registered images. Analysis was also performed on adjacent sections adding the effect of repeated measurement. The mean values for intensities (PSL/mm2 values) were calculated.

Histology

After autoradiography, the sections were stained in haematoxylin and eosin and studied for morphology. The degree of inflammation in the whole aortic tissue was semi-quantitatively assessed by an experienced pathologist as the number of leucocytes in the tissue (no leucocytes = non-inflamed aorta; occasional single polymorphonuclear leucocytes or lymphocytes = mild inflammation; occasional polymorphonuclear leucocytes or lymphocytes and some groups of inflammatory cells = moderate inflammation; abundant infiltration of polymorphonuclear leucocytes and/or lymphocytes = severe inflammation) [4]. For ROI analysis, each plaque area was defined according to the histology as non-inflamed (none or occasional leucocytes) or inflamed (moderate or severe inflammation).

Adjacent sections were immunostained with Mac-3 (antibody for mouse macrophages) and PBR antibody. Eight micrometer sections stored at −70°C were melted, air-dried and fixed in ice-cold acetone. Endogenous peroxidase was blocked and the sections were incubated for 1 h with Mac-3 antibody. Control immunostaining was performed with IgG-negative control antibody. Biotinylated secondary antibody was incubated for 30 min, followed by incubations with streptavidin–horseradish peroxidase and AEC Substrate Chromogen according to the instructions of the manufacturer. Finally, the sections were counterstained in haematoxylin. All antibodies were from BD Pharmingen and all reagents were from DakoCytomation, CA, USA.

Statistical methods

All results are expressed as the mean±SD. Univariate correlations were calculated using the Spearman correlation method. Analysis of variance (ANOVA) was used to study the significance of differences observed in animal characteristics and biodistribution between LDLR/ApoB48 mice and control animal groups. In autoradiography analysis, parametric ANOVA of repeated measures was used in the inter-animal comparison for the mean PSL values of radioactive uptake of defined ROI. For intra-animal comparisons, ANOVA was used. A p value less than 0.05 was considered as statistically significant.

Results

Characterisation of atherosclerosis in LDLR/ApoB48 mice

All LDLR/ApoB48 mice had extensive atherosclerosis showing plaques with different characteristics throughout the aortic tree. The observed plaques in the aorta contained cell-rich, inflamed areas and acellular necrotic cores, mostly resembling type V lesions [22]. Leucocyte infiltration was observed in plaques in all mice and, based on the histological analysis, the inflammation was estimated to be of a moderate degree (occasional polymorphonuclear leucocytes or lymphocytes and some groups of inflammatory cells). Macrophage staining revealed areas in the plaques with moderate number of macrophages and PBR-specific antibody stained mainly the same sites as the macrophage antibody (Fig. 1). In control mice aortas, no atherosclerosis was detected.

Fig. 1
figure 1

A plaque in the brachiocephalic artery trunk, concecutive sections. a Macrophage staining (Mac-3) and b PBR-specific antibody. L lumen, P plaque, W wall. Original magnification ×100

Full size image

Binding of [3H]PK11195 in aortic sections

The presence of PK11195 binding sites in the atherosclerotic plaques was verified by examining the in vitro competive binding of tritium-labelled PK11195 onto aortic sections (41 sections from five mice). [3H]PK11195 was found to bind to the aortic sections in the sites of the plaques and also on the healthy wall (Fig. 2). The specific vs. non-specific binding ratio was high (12-fold).

Fig. 2
figure 2

a Contour image of a section of atherosclerotic mouse aorta; b in vitro binding of [3H]PK11195; c adjacent section, in vitro [3H]PK11195 with a blocking dose of cold PK11195; d–f representative sections, HE stain. Bar = 0.4 mm

Full size image

Aortic uptake and biodistribution of [11C]PK11195 in LDLR/ApoB48 and control mice

The uptake of [11C]PK11195 in the whole aorta of atherosclerotic LDLR/ApoB48 mice (4.96 ± 0.16%ID/g, n = 5) was significantly lower than that in the control mice (9.00 ± 0.12%ID/g, n = 3) (p = 0.004) (Table 1). In plasma, lung and heart, the [11C]PK11195 uptake did not differ between the controls and the LDLR/ApoB48 mice. The availability of [11C]PK11195 in the circulating blood was not found to be affected by the animal weight (p = 0.26). The uptake ratio of aorta to plasma was 22 ± 7 and aorta to heart was 0.6 ± 0.3 in LDLR/ApoB48 mice.

Table 1 The biodistribution [%ID/g (mean ± SD)] of intravenously administered [11C]PK11195 in atherosclerotic LDLR/ApoB48 mice and control C57Bl/6N mice
Full size table

Distribution of [11C]PK11195 in aortic sections

In the longitudinal sections of the aorta, both healthy vessel wall and atherosclerotic plaques were found. Autoradiograms of nine to 12 sections of each animal were analysed. In total, 136 non-inflamed plaque regions, 279 inflamed plaque regions and 147 healthy vessel wall regions were analysed from autoradiograms (Figs. 3 and 4). In each mouse, the mean uptake of [11C]-radioactivity into each region were calculated (Table 2).

Fig. 3
figure 3

Example of autoradiography analysis (mouse no. 108)

Full size image
Fig. 4
figure 4

An example of analysis of the aortic sections. a HE stain of the section, b autoradiography image with superimposed contour image, c ARG image, d healthy vessel wall, e non-inflamed plaque region, f inflamed plaque region. Original magnification ×200 (d–f), scale bar = 100 μm

Full size image
Table 2 Autoradiography analysis of accumulated radioactivity into vascular areas in LDLR/ApoB48 mice. The intensity (PSL/mm2, mean ± SD)
Full size table

Autoradiography analysis revealed diffuse [11C]PK11195 uptake in the aorta sections, which was not specifically limited to the inflamed plaques. There was significantly higher accumulation of 11C radioactivity to the inflamed regions compared to the non-inflamed regions of the plaques (p = 0.011), but no difference in the accumulation in the inflamed regions and the healthy vessel wall of the same animals was found. The uptake to cell-rich sites in plaques did not differ from the uptake to the healthy vessel wall (p = 0.1) (Table 2).

Discussion

The aim of this study was to investigate the uptake of [11C]PK11195 in atherosclerotic plaques of the LDLR/ApoB48 mice and estimate its potential to be used in the imaging of inflammation associated with atherosclerosis. We found that the uptake of tracer was higher in the inflamed atherosclerotic plaques than in the non-inflamed plaques. However, the tracer uptake to other structures of the artery wall, to both healthy wall and adventitial tissue was also evident and comparable to the uptake into the inflamed plaques. This suggests that clinical imaging of vascular inflammatory processes with [11C]PK11195 is challenging and a tracer with a better target-to-background ratio would be likely needed.

The presence of PK11195 binding sites in the atherosclerotic plaques was verified in an in vitro setting. In the displacement studies, the binding of [3H]PK11195 was found to be specific. However, already in these studies, [3H]PK11195 was found to bind to the healthy artery wall in addition to plaques. In the subsequent ex vivo studies, also quite diffuse [11C]PK11195 uptake was seen, not specifically limited to the inflamed plaques. This finding was supported by the results that the uptake of [11C]PK11195 (%ID/g) was not higher in the aorta of atherosclerotic mice compared to the control mice.

PK11195 is a high-affinity ligand to PBR. PBR, also known as translocator protein [8], is present in the mitochondrial outer membrane as a part of the complex formed by voltage-dependent anion channel and adenine nucleotide transporter [23]. Relatively high expression of PBR can be found in leucocytes especially in monocyte/macrophages and neutrophils but it is also expressed in adrenal gland, lung, kidney and heart, and in other tissues in lesser amounts [10, 11, 24]. In the brain, PBR can only be found in low numbers in the glial cells and astrocytes, but in neurodegenerative states, the activated glial cells express high amounts of PBR [25]. [11C]PK11195 has previously been used for PET imaging of inflammation related to multiple sclerosis [21], and it has also been suggested as a suitable agent for the imaging of atherosclerosis-associated inflammation. The accumulation in the inflamed lung in animal models and humans [15–17] have been shown to be significantly increased due to activated macrophages in the inflamed area.

It is possible that macrophages may express PBR differently in response to their environment. A very recent report by Branley et al. shows that, although [3H]PK11195 binds solely to macrophages of bronchoalveolar lavage cells in patients with interstitial lung disease (ILS), the total binding in the lung is diminished compared to healthy controls [26]. They suggest that, in ILS and in other inflammatory diseases like osteoarthritis [27], the reduced expression of PBR in the macrophages may be a reflection of a proapoptotic and proinflammatory milieu. We do not know the amount of PBR specifically in the macrophages in the atherosclerotic plaques, but it is possible that although the macrophages are present in the plaques, the level of PBR expression may differ as a reflection of their altered functionality and environment.

In this study, the uptake of [11C]PK11195 in the whole aorta of atherosclerotic LDLR/ApoB48 mice was found to be lower compared to the aortas of the C57Bl control mice. This could be explained by the high number and mass of core areas of the plaques observed in the aorta of atherosclerotic mice. As autoradiography analysis showed, [11C]PK11195 does not accumulate to these non-inflamed sites. The atherosclerotic aortas have higher mass due to these plaques, and therefore, the radioactivity calculated against the weight per gram tissue is lower when accumulation in other sites is similar.

Limitations

The biodistribution studies were performed at one time point of 20 min. [11C]PK11195 has been found to metabolise rapidly in vivo in mice and humans [28]. According to our studies, 20 min after injection, the amount of unchanged [11C]PK11195 was 78% in human plasma and 85% in rat plasma [29]. In this study, we chose to use a time point of 20 min due to the short half-life of 11C and detection difficulties of samples of small size. With a longer time point, the target-to-background ratio might have been better, although at the time point of 20 min, the tissue-to-blood ratios were already very good.

The physiological differences between the controls and atherosclerotic mice may explain the observed variance in the uptake of [11C]PK11195 in the aorta. In the biodistribution study, the animal weight was not considered to be a source of bias since the availability of [11C]PK11195 in the circulating blood was not affected by the animal weight (p = 0.26). However, there was a significant difference in the uptake of [11C]PK11195 in the liver. LDLR/ApoB48 mice could suffer from compromised liver functions due to atherosclerosis and high plasma lipid levels, which could affect the uptake in the liver. As for the other studied tissues (plasma, erythrocytes, lung and heart), the [11C]PK11195 uptake did not differ between the controls and the LDLR/ApoB48 mice.

The variance in the ROI analysis may partly be due to the small size of the analysed areas, especially in the healthy vessel wall, which is only a few cell layers in thickness. Interference from other nearby tissue compartments was, however, diminished by drawing ROIs for the healthy wall region farther away from the plaques. The variance observed in the plaque ROIs may be partly due to the heterogenity of cellular composition in the 20-μm sections. Macrophages were detected in the plaques by comparing the HE-stained autoradiography sections to consecutive macrophage-stained sections and directly from HE-stained sections by morphology. The inflammatory activity was defined based on the information of the consecutive macrophage-stained section and cellularity of the site in these advanced plaques and the macrophages were not specifically counted, therefore, these sites may contain also other leucocytes and occasional smooth muscle cells. However, for clinical application, it is more important to differentiate the active, cell-rich, inflamed plaques from non-inflamed plaques, since these are pathologically very different and their vulnerability prognosis is essentially different.

Conclusions

Our results indicate that the uptake of [11C]PK11195 to the atherosclerotic vessels is not limited to inflamed plaques. In a mouse model of atherosclerosis, the uptake to inflamed regions in plaques was more prominent than to non-inflamed areas, but did not differ from the uptake to the healthy vessel wall. This binding to surrounding tissue resulting in high background may limit the usability of [11C]PK11195 in clinical imaging of atherosclerotic plaques.