Human carotid artery specimens were obtained during surgery from patients undergoing carotid artery endatherectomy for significant stenosis after informed consent. The study was approved by the institutional ethical committee (Reg. No.: 4IRutsch1). Immediately after excision of the carotid atheromas, the samples from the arteria carotis interna were split longitudinally in two samples, one was directly frozen in liquid nitrogen, the other was immediately embedded in cryoprotective medium (Thermo Shandon, Frankfurt/Main, Germany) and snap-frozen in liquid nitrogen for subsequent cryosectioning. Normal human vascular tissue including samples with minimal intima hyperplasia (Aorta, N= 16; A. mammaria, N= 23; A. radialis; N= 11) was obtained from patients undergoing coronary bypass surgery (approved by the ethics committee of Muenster University Hospital Az. 2007-210-f-S). The frozen samples were stored at −80°C until required.

TRIzol extraction of total RNA was carried out essentially according to the manufacturer’s specifications (Invitrogen, Carlsbad, CA, USA) with minor modifications. To a carotid artery segment, 1.6 ml of TRIzol was added, and the sample was homogenized using a Mikro-Dismembrator U (B. Braun Biotech International, Melsungen, Germany). Following incubation for 5 min. at room temperature, 0.3 ml of chloroform was added, and the tubes were shaken for 20 sec. and then incubated for 3 min. at room temperature. This mixture was then centrifuged at 12,000 x g for 15 min. at 48C. The aqueous phase was transferred to a new tube, 500 ml TRIzol and 100 ml chloroform were added, the tubes were shaken for 20 sec. and centrifuged at 12,000 x g for 15 min. at 48C. The aqueous phase was transferred to a new tube, isopropanol (0.8 ml) was added and the sample was mixed by inversion and then incubated for 10 min. at room temperature prior to centrifugation at 12,000 x g for 20 min. at 48C. The RNA pellet was washed with 1 ml of 80% ethanol, centrifuged at 7500 x g for 15 min. at room temperature, and air-dried in sterile conditions for 5–10 min. The RNA pellets were dissolved in 20 ml of RNase-free water and stored at −808C. RNA quality and integrity was determined by analysing the absorbance at 260/280 nm wavelength and by visualizing the 28S and 18S rRNA bands in ethidium bromide-stained agarose gels under ultraviolet light. The A260/A280 ratio of the total RNA samples was >1.7 and 28S/18S ratios were >1.8.

Total RNA isolated from carotid endatherectomy samples was used. One microgram of total RNA was reverse-transcribed after an initial denaturation step at 658C for 10 min. in a total volume of 20 ml using 10 U/ml of Superscript™ III reverse transcriptase (Invitrogen™ Life Technologies), 1 U/ml Ribo Lock™ ribonuclease inhibitor (MBI Fermentas GmbH, St. Leon-Rot, FRG), 25 ng/ml oligo[(dT)12–18] (Invitrogen™ Life Technologies), 0.5 mmol/l each of 29-desoxynucleoside-59triphosphate (dNTP), 5 mmol/l dithiothreitol (DTT), 1x first strand buffer (250 mmol/l Tris-HCl (pH 8.4), 375 mmol/l KCl, 15 mmol/l MgCl2) freshly diluted from 5x stock (Invitrogen™ Life Technologies) at 378C for 60 min. Samples were heated at 958C for 5 min. to terminate RT activity. Total RT products were subsequently used for PCR amplification. mRNAs of NPP1, OPN and glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) were assessed in all biopsies by RT-PCR. In brief, 1 mg total RNA was reverse transcribed using Superscript™ III according to the manufacturer’s instructions (Invitrogen™ Life Technologies). The RT products (1ml) were brought to a volume of 25 ml containing 1x Qiagen PCR buffer, 2.5 mM MgCl2, 0.1 mM of each dNTP, 1.5 U of Taq Polymerase (Qiagen GmbH, Hilden, Germany) and 0.4 mM of each primer. Primers were designed across intron/exon boundaries. Optimal amplification temperature and cycle number were estimated on a gradient cycler (Stratagene, Cedar Creek, TX, USA). After estimating the optimal temperature (50–608C, 28C steps) a cycle gradient was performed (24–39 cycles, 3-cycle steps) using the optimal temperature. The optimal cycle number was in the range of linear amplification. Table S1 summarizes the individual sequences of the primers, annealing temperatures and cycle numbers used for each primer pair and the PCR product sizes (Table S1). After denaturation at 958C for 5 min. amplification was carried out in a Robocycler Gradient 96 (Stratagene) using the following temperature and time profile: denaturation at 958C for 30 sec.; primer annealing at the respective temperatures for 60 sec.; primer extension at 728C for 60 sec.; and a final extension of 728C for 10 min. A total of 10 ml of the PCR reaction were analysed using standard agarose gel electrophoresis. To evaluate the relative expression, the agarose gels were scanned using a video documentation system, and ImageQuant™ software (Molecular Dynamics, Krefeld, Germany). The absorption units were corrected for GAPDH mRNA levels.

NPP1, OPN and SMC specific a-actin were detected as follows: (i) for NPP1, monoclonal mouse anti-human NPP1 1:500 (3E8, 0.25 mg/ml IgG, kindly provided by James Goding, Monash University, Prahan, Australia) [19]; (ii) for OPN, monoclonal mouse anti-human OPN 1:100 (OP3N, 0.34 mg/ml IgG, Novocastra Laboratories Ltd., Newcastle upon Tyne, UK) (iii) SMC specific a-actin, monoclonal mouse anti-human SMC a-actin 1:100 (HHF35, 0.11 mg/ml IgG, Dako Corporation, Carpinteria, CA, USA) were used. The secondary antibody detection system was donkey anti-mouse conjugated to Cy3 (1:50; 0.5 mg/ml IgG, Chemicon International, Temecula, CA, USA). Nuclei were stained with 4'-6-diamidino-2-phenylindole (DAPI, 1:10000; Carl Roth GmbH & Co.KG, Karlsruhe, Germany).

Cryosections were fixed for 20 min. in methanol at 4°C and blocked for 1 hr in 5% normal donkey serum containing 1% bovine serum albumin (BSA) at room temperature. Sections were incubated with the respective primary antibody diluted in PBS containing 5% normal donkey serum and 1% BSA at room temperature over night. After washing thoroughly in PBS, sections were treated with the secondary antibody diluted in PBS containing 1% BSA for 1 hr at room temperature. After rinsing with PBS, nuclei were stained for 5 min. by DAPI. After washing again, the slides were mounted with DakoCytomation fluorescent mounting medium (Dako, Hamburg, Germany). Negative controls included substitution of the primary antibody by mouse anti-human IgG 2a (1:500; MOPC-173, 0.25 mg/ml, kindly provided by James Goding, Monash University, Prahan, Australia) or omission of the primary antibody. Results were documented using a fluorescence microscope (AXIOPHOT2™; Zeiss, Jena, Germany) fitted with the appropriate filter blocks for detection of Alexa and Texas Red fluorescence.

For visualization of calcification the von Kossa staining procedure was used. Sections (10 μm) were stained according to standard procedures. Haematoxylin and eosin standard stain demonstrated general morphology.

For quantification of calcification the frozen specimens were scanned using a Dual Source CT (Somatom Definition, Siemens Medical Solutions, Forchheim, Germany). Scan parameters were as follows: detector collimation: 2 × 64 × 0.6 mm, pitch: 0.7, 330 msec. rotation time, tube current-time product: 400 mAs and tube voltage 120 kV. We used a simulated ecg-signal (60 bpm) and images were reconstructed with a slice thickness of 3.0 mm and an increment of 1.5 mm. The field of view was set to 250 mm. These parameters were chosen as the scanner is calibrated to these settings using a calcium score phantom (QRM, Moehrendorf, Germany).

As the scan protocol is optimized for coronary calcium quantification, we used a simulated ecg-signal (60 bpm). Images were reconstructed with a slice thickness of 3.0 mm and an increment of 1.5 mm. The field of view was set to 250 mm. These parameters were chosen as the scanner is calibrated to these settings using a calcium score phantom (QRM).

The quantification of the calcium mass was performed with standard software (CalciumScore, Siemens Medical Solutions) running on the scanner. To calculate the calibrated calcium mass, the density and volume of each volume element (voxel) with a density above 130 Hounsfield units are taken into account [20].

Expression data are presented as means ± S.D. The groups were compared with ANOVA followed by post hoc analysis of the means. Univariate analysis was performed by means of the Mann-Whitney U-test for continuous variables. Fisher’s exact test was used to examine the significance of the association between two variables. A value of P < 0.05 was considered as statistically significant.

Forty-five patients (average age 67 ± 8.3 years) were enrolled in the study, including 37 males and 8 females. In our study cohort, 7 patients (15.6%) showed a non-calcified plaque (–calc group), 38 (84.4%) showed various degrees of calcification (+calc group). Both groups did not differ with respect to total cholesterol (calcified: 214.20 ± 102.26 versus non-calcified: 214.52 ± 34.20 mg/dl, n.s.), LDL (calcified: 139.84 ± 26.73 versus non-calcified: 167.70 ± 46.81 mg/dl, n.s.), HDL (calcified: 56.92 ± 16.57 versus non-calcified: 59.60 ± 18.98 mg/dl, n.s.) and triglyceride levels (calcified: 231.00 ± 140.13 versus non-calcified: 437.80 ± 343.36 mg/dl, n.s.). Table S2 summarizes the clinical characteristics of the patients.

Calcification levels as shown by CT analysis varied between 0.01% and 58.9% of plaque mass. Linear regression analysis (Fig. 1) of all samples with known extent of calcification, which yielded intact RNA (N= 45) revealed a positive correlation between age and calcification (R= 0.44, P < 0.01). Correlations of calcification with the other risk factors of atherosclerosis (diabetes mellitus, smoking, obesity, lipid parameters, gender) were not found.

The mRNA expression levels of NPP1 and OPN were assessed by RT-PCR in normal arteries (all not calcified, including samples with minimal intima hyperplasia) and all carotid artery tissue samples yielding intact RNA (N= 34). The mRNAs of NPP1 and OPN were detected in all samples.

The NPP1 mRNA levels ranged from 4.9 to 13.8 in normal arteries. In carotid endatherectomy samples the levels of NPP1 mRNA ranged from 3.9 to 10.7 in non-calcified atheroma and from 1.2 to 5.6 in calcified samples. Differences between normal arteries and non-calcified carotids were not found (normal: 9.51 ± 4.87 versus non-calcified: 7.62 ± 2.82, P= n.s.). For analysis of mRNA expression, calcified samples were divided into moderately and severely calcified samples. Samples showing calcification up to 5% of the plaque mass were considered as moderately calcified (+calc), samples showing calcification of more than 5% were considered as severely calcified (++calc). In comparison with normal arteries, the NPP1 mRNA levels of moderately and severely calcified carotids were significantly decreased (normal: 9.51 ± 4.87 versus calcified: +calc: 1.90 ± 0.82, P < 0.0001; normal: 9.51 ± 4.87 versus++calc: 3.62 ± 1.25, P < 0.0001). Also, in comparison with non-calcified atheroma, NPP1 mRNA levels were significantly decreased in moderately and severely calcified samples ([non-calcified: 7.62 ± 2.82 versus calcified: +calc: 1.90 ± 0.82 and ++calc: 3.62 ± 1.25, P < 0.0001]; Fig. 2).

In normal arteries, OPN mRNA levels ranged from 0.6 to 3.4. In carotid endatherectomies OPN mRNA expression levels were between 2.7 and 9.2 in non-calcified and between 1.1 and 6.4 in calcified samples. OPN levels were significantly up-regulated in non-calcified atheromatous carotid samples compared to normal arteries (normal: 1.88 ± 0.96 versus non-calcified: 5.32 ± 2.81, P= 0.0001), and in calcified carotid samples compared to normal arteries (normal: 1.88 ± 0.96 versus calcified: +calc: 2.56 ± 1.06, P < 0.05 and normal: 1.88 ± 0.96 versus calcified: ++calc: 3.47 ± 1.40, P < 0.0001). When comparing the non-calcified and calcified atheroma groups, lower levels of OPN were found in both moderately and severely calcified samples ([non-calcified: 5.32 ± 2.81 versus calcified: +calc: 2.56 ± 1.06, P > 0.0001 and non-calcified: 5.32 ± 2.81 versus++calc: 3.47 ± 1.40, P < 0.05]; Fig. 2). Similar to NPP1 expression levels, the lowest levels of OPN mRNA were observed in moderately calcified atheromatous carotid samples (Fig. 2).

Linear regression analysis was used to establish the potential relationship between NPP1 and OPN as well as the relation between NPP1, OPN and calcification. In the –calc group correlations between NPP1, OPN and age were not found. In the +calc group, age positively correlated with calcification (R= 0.38483, P= 0.047). The mRNA expression levels of NPP1 as well as OPN positively correlated with the degree of calcification (NPP1: R= 0.53975, P= 0.00019; OPN: R= 0.46, P= 0.036), but not with age. Furthermore, mRNA levels of NPP1 and OPN positively correlated with each other (R= 0.35, P= 0.021; Fig. 3).

In the +calc group, NPP1 and OPN mRNA levels in the plaque were analysed with respect to gender, presence of symptoms, hypertension, smoking, obesity, diabetes and therapy with statins (Table 1). NPP1 levels were lower in male (61%; P= 0.043) as well as in obese patients (54%; P= 0.003), whereas increased levels were found in patients with hypertension (188%; P < 0.0001). Decreased OPN mRNA levels were found in patients with hypercholesterolemia (80%; P= 0.049) and diabetes (81%; P= 0.013); the levels were increased in the smoking subgroup (136%; P= 0.023).

Under statin therapy levels of NPP1 and OPN significantly decreased to about 70% (P= 0.009 and P = 0.014, respectively). Moreover, the extent of calcification was significantly lower in statin treated patients of the +calc group (+statin: 4.04 ± 4.31 versus–statin: 10.79 ± 9.98, P = 0.044).

Sections of plaque samples were analysed by immunofluorescence with respect to the expression and distribution of NPP1 and OPN in different plaques with various degrees of calcification (Figs 4–6).

In non-calcified plaques NPP1 was detected mainly in the media. In the non-calcified core region of a plaque (Fig. 4), few NPP1 expressing cells were found (Fig. 4E, F). In moderately calcified areas of the plaque (Fig. 5) NPP1 was detected in the media and in intimal regions adjacent to the media (Fig. 5E, F). In the media NPP1 colocalized with smooth muscle cells (Fig. 5G, H). In the severely calcified plaque (Fig. 6) NPP1 was found in the acellular matrix of the non-calcified core area with only few expressing cells (Fig. 6E, F). OPN immunolabel gradually increased with progressing plaque calcification. In the uncalcified plaque (Fig. 4I, J) no OPN expressing cells were detected. In moderately calcified plaques, a positive staining was found in the intimal layer adjacent to the calcification (Fig. 5I, J). OPN was particularly strongly expressed in the acellular severely calcified areas of the plaque core (Fig. 6I, J).

As shown in the present study, age correlated with the degree of calcification of atheromatous carotids in patients with proven calcification of the atheroma. This is in line with previous studies showing that, in patients with coronary heart disease and atherosclerosis, the extent of calcification highly correlates with age and mortality [7]. In carotid plaques, destabilization and thus fatal ischemic events have been associated with increased deposition of calcified matrix vesicles and microcalcifications [21]. In patients with end-stage renal disease the coronary calcium score increased with age [7]. Also, the presence of calcium in renal arteries was significantly associated with age [22]. To our knowledge, our investigations represent the first study confirming this association also in carotid atheromatous plaques.

As shown for the first time in the present study NPP1 – an enzyme generating the mineralization inhibitor PPi – is down-regulated in calcified stenotic carotid arteries, i.e. in intimal calcification. In normal arteries and in non-calcified atherosclerotic plaques, NPP1 expression levels are comparable. As previously reported, NPP1 plays an important role in medial calcification in human beings and animal models [16, 23, 24]. NPP1 levels and activity are reduced or even deficient in patients with generalized arterial calcification of infancy, a disorder caused by mutations in the ENPP1 gene [16, 23]. Moreover, ENPP1 polymorphisms have been associated with insulin resistance and atherogenic phenotypes, including coronary calcification and myocardial infarction [17, 18, 25]. The pathogenic mechanism underlying this link is most likely to be on the one hand a consequence of systemic abnormalities, which are associated to insulin resistance, such as arterial hypertension, obesity and dyslipidaemia [26]. On the other hand, atherosclerosis may derive secondary to a direct deleterious effect of impaired insulin signalling on the endothelium leading to reduced insulin-stimulated nitric oxide release and decreased vasodilatation [27]. However, these systemic effects, which are associated with increased NPP1 plasma levels interfering with insulin signalling [28], were beyond the scope of our study.

Interestingly, although generally down-regulated in calcified carotid atheroma, NPP1-levels positively correlated with the degree of calcification in our study. Our data suggest that a reduction of NPP1 expression is also a trigger of intimal calcification. The positive correlation of NPP1 levels with the degree of plaque calcifications furthermore suggests a counter-regulative mechanism, which in the end turns out to be insufficient to prevent further progression of calcification. Such a regulatory role has already been proposed for other inhibitory matrix proteins, including matrix Gla protein (MGP) [29] and OPN [29, 30], the expression of which has been found to correlate with increasing plaque calcification.

OPN mRNA levels were also up-regulated in carotid plaques in our study. However, marked deposits of OPN were found only in strongly calcified carotid atheromatous lesions closely associated with hydroxyapatite clusters. OPN is a known inhibitor of vascular mineralization [31–33]. Typically, OPN is associated with mineralized deposits in human atherosclerotic plaques [33]. Reportedly, OPN serum levels are elevated in patients with carotid stenosis and have been discussed as potential markers of vulnerable carotid plaques [34]. Taken together these data and ours, one may suggest that during atherogenesis and early stages of calcification, OPN is released into the circulation. Only with progressing calcification OPN is deposited and accumulated in the calcified area of the plaque.

Interestingly, in calcified carotid plaques, we observed a positive correlation between NPP1 and OPN expression levels, indicating a common regulatory pathway. This would be in line with previous studies suggesting down-regulated OPN expression in defective NPP1 expression [35]. The results of our study furthermore prove the current concept of arterial calcification being an active process, which is orchestrated by accelerators of calcification including alkaline phosphatase, BMP-2, BMP-4, osteocalcin and bone sialoprotein and by inhibitors such as MGP and OPN [29, 30].

Differential relations between NPP1, OPN and pro-atherogenic factors were observed in the present study.

Lower levels of NPP1 were detected in the plaque of male and in obese patients; both groups are known to have an increased prevalence of atherosclerosis and increased burden of calcification. As shown by Gonzales-Sanchez and coworkers the ENPP1 121Q allele might contribute to the genetic susceptibility to abdominal obesity among subjects with metabolic syndrome [36]. Moreover, adipocyte maturation is enhanced in the absence of NPP1, and adipocyte maturation was more efficient in mesenchymal embryonal cells from ENPP1 knockout mice than from wild-type mice [37]. Our data indicate that in both groups compromised inhibition of calcification by NPP1 may augment the intimal calcification. However, we did not measure NPP1 serum levels in the patients to validate this hypothesis.

OPN expression was lower in calcified plaques of patients with hypercholesterolemia and diabetes. Reportedly, both hypercholesterolemia and diabetes are pro-atherogenic and are directly related to the prevalence and extent of coronary artery calcification [38, 39]. Plasma levels of OPN are not associated with plasma cholesterol metabolites but with tissue markers of cholesterol synthesis, indicating an interplay between OPN and cholesterol metabolism in the plaque [40].

Interestingly, in our study lipid lowering therapy by statins was associated with decreased expression levels of NPP1 and OPN in the plaque. Furthermore, plaques of statin treated patients were less calcified. Reportedly, statins not only reduce the anatomic burden of coronary artery disease but also that of calcification [41, 42]. Whether the effect of statins on NPP1 and OPN is mediated via their effect on lipid metabolism or by a pleitropic mechanism has to be elucidated.

As mentioned above, expression of NPP1 and OPN positively correlated in the calcified carotid plaque. Since both inhibitors of calcification are affected by alterations of the lipid metabolism and lipid profile, looking for a common regulating pathway in this area might be promising.

Taken together, our study demonstrates that the inhibitor of media calcification NPP1 is down-regulated in calcified advanced plaques. Interestingly, NPP1 levels are higher in severely calcified lesions, most likely reflecting a counter-active mechanism. We therefore conclude that NPP1 serves as a novel player orchestrating plaque calcification. Moreover, our data suggest an interplay between NPP1, OPN and potent risk factors of atherosclerosis.