Despite an only minor reduction in the glomerular filtration rate, uninephrectomy (UNX) markedly accelerates the rate of growth of atherosclerotic plaques in ApoE−/− mice. It has been suggested that vitamin D receptor (VDR) activation exerts an antiproliferative effect on vascular smooth muscle cells, but the side effects may limit its use. To assess a potentially different spectrum of actions, we compared the effects of paricalcitol and calcitriol on remodeling and calcification of the aortic wall in sham-operated and UNX ApoE−/− mice on a diet with normal cholesterol content. Sham-operated and UNX mice were randomly allotted to treatment with solvent, calcitriol (0.03 μg/kg) or paricalcitol (0.1 μg/kg) 5 times/wk intraperitoneally for 10 wk. Semithin (0.6 μm) sections of the aorta were analyzed by 1) morphometry, 2) immunohistochemistry, and 3) Western blotting of key proteins involved in vascular calcification and growth. Compared with sham-operated animals (5.6 ± 0.24), the wall-to-lumen ratio (x100) of the aorta was significantly higher in solvent- and calcitriol-treated UNX animals (6.64 ± 0.27 and 7.17 ± 0.81, respectively, P < 0.05), but not in paricalcitol-treated UNX (6.1 5 ± 0.32). Similar differences were seen with respect to maximal plaque height. Expression of transforming growth factor (TGF)-β1 in aortic intima/plaque was also significantly higher in UNX solvent and UNX calcitriol compared with sham-operated and UNX paricalcitol animals. Treatment with both paricalcitol and calcitriol caused significant elevation of VDR expression in the aorta. While at the dose employed paricalcitol significantly reduced TGF-β expression in plaques, calcitriol in contrast caused significant vascular calcification and elevated expression of related proteins (BMP2, RANKL, and Runx2).
- vitamin D/analogs and derivates
- vascular calcification
recently, cardiovascular disease in the early stages of chronic kidney disease (CKD) has attracted considerable attention, and reduced renal function has been recognized as an independent risk factor for cardiovascular mortality (1, 12, 15, 21). Atherogenesis is accelerated by CKD, most probably due to the chronic inflammatory state of this condition (10). Indeed, higher expression of inflammatory markers in atherosclerotic plaques and more intense calcification of the media are found in the coronaries of uremic patients compared with matched nonrenal patients with coronary disease (16).
Although in dialyzed patients vascular calcification had been documented in the distant past (8, 26), the interest in this problem has been rekindled after the observation that high predialysis phosphate concentrations increase not only overall mortality (3) but specifically also cardiovascular mortality (13).
In a retrospective observational study, Teng et al. (33) found that all cause and cardiovascular mortality were lower in dialysis patients treated with paricalcitol compared with patients treated with calcitriol (33). Controlled prospective evidence is still lacking, however (9). Vitamin D receptor (VDR) activators have been reported to upregulate the calcification inhibitor matrix gla protein (MGP) (11) and the synthesis of vasoprotective prostacyclin (34). They also increase the vasoprotective activity of estrogens (29, 30). These observations do not provide a coherent pathomechanistic explanation of the mechanisms underlying the effect of VDR activation on vessels.
This prompted us to compare the effect of two VDR activators, i.e., paricalcitol and calcitriol, in an established model of atherogenesis, the ApoE knockout mouse with and without partial renal ablation.
MATERIALS AND METHODS
Ninety-six male ApoE−/− mice (12 wk of age, Charles River, Brussels, Belgium) were fed a standard rodent diet (19.0% protein, 3.3% fat, 1.0% calcium, and 0.7% phosphorus) provided ad libitum.
After a 15-day adaptation period, mice were submitted to uninephrectomy (UNX; n = 48) or sham operation (n = 48) under isoflurane anesthesia (Isoflurane, Baxter). For UNX, the right kidney was removed, and for the sham operation the kidney was decapsulated.
Two weeks after surgery, animals were randomly assigned to the following 10-wk treatments (administered intraperitoneally): 1) untreated: solvent alone (50% water, 20% alcohol, and 30% propylene glycol); 2) paricalcitol: 19-Nor-1α,25-dihydroxyvitamin D2 (Abbott, Chicago, IL) at a dose of 0.1 μg/kg five times per week; and 3) calcitriol: 1-α,25-dihydroxyvitamin D3 (Abbott) at a dose of 0.03 μg/kg five times per week. The drug solutions were prepared in a laminar flux chamber, and drug concentrations were calculated using spectrophotometry (ND 1000, NanoDrop Tech, Wilmington, DE).
Systolic blood pressure assessment (tail plethysmography) and 24-h urine collection (metabolic cages) were performed in the last week of the experiment. Twelve weeks after UNX, mice were euthanized and perfused under nonpulsatile 90- to 110-mmHg pressure with either 3% glutaraldehyde for morphological analysis and the next series of animals with 4°C NaCl for immunohistochemical and Western blot analysis. Blood was obtained just before the infusion. All procedures were approved by the local ethics committee for animal experiments (Regierungspräsidium, Karlsruhe, Germany).
Proteinuria was measured by ELISA using a specific kit (Mouse Albumin Elisa Quantitation Kit, Bethyl Labs, Montgomery, TX), and blood samples were freshly analyzed for ionized calcium by an Ionometer (Fresenius EH-F, Bad Homburg, Germany). Analysis of phosphorus, creatinine, and cholesterol was performed by photometry (Labor Limbach, Heidelberg, Germany). Because of the limited size of serum samples, lipoprotein analysis could be performed only in eight sham-operated solvent, five sham-operated paricalcitol, five sham-operated calcitriol, eight UNX solvent, and four UNX paricalcitol and four UNX calcitriol animals.
The thoracic aorta of each animal was carefully taken out and divided into two parts. In glutaraldehyde-perfused animals, the first 5 mm of the aorta were fixed in epon-araldite. Semithin 0.6-μm sections were stained with 0.1% toluidine blue solution. The remaining aorta was embedded in paraffin and cut (4-μm sections), which were stained with Elastica van Gieson (EVG), Sirius red, and von Kossa. At least 25 sections of the aorta per animal were made from 4 standardized segments of the aorta. In 0.9% NaCl-perfused animals, the upper half of the aorta was fixed in 4% phosphate-buffered formaldehyde and embedded in paraffin for histology and immunohistochemistry. The lower half of the obtained tissue was snap frozen in liquid nitrogen and stored at −86°C until analysis.
Morphological and histological evaluation.
All investigations were performed in a blinded manner; i.e., the observer was unaware of the study group. Images of the aortic rings were obtained with a digital camera (Polaroid DMC 1, Polaroid, Cambridge, MA) attached to an optical microscope (Olympus BX45, Tokyo, Japan) under different magnifications (×40–200). Images were analyzed using automatic image-analysis system (Optimas 6.0, Seattle, WA) software. Mean wall thickness, lumen, intima, and media layer diameters were determined in every artery using semithin sections.
Sirius red and von Kossa stainings were performed in all glutaraldehyde- and NaCl-perfused animals (n = 16/study group). Sirius red staining was analyzed using image software (Image Pro Plus 5.0, Media Cybernetics, Bethesda, MD). The results are given as percent stained area/area of interest. Von Kossa staining was analyzed using a semiquantitative scoring system (0–4), which was rated according to the area stained (0, none; 1, <10%; 2, 10–50%; 3, 51–75%; and 4, >75%).
Immunohistochemical staining was performed in 4-μm-thick paraffin sections mounted on positive-charged slides (Superfrost plus, Microm, Walldorf, Germany). Sections were deparaffinized with xylene and rehydrated through descending concentrations of ethanol. Antigen retrieval was performed by heating the slides in Target Retrieval Solution, pH 9.0 (Dako Denmark, Glostrup, Denmark) at 97°C in a water quench for 5 min for the interstitial and cell surface antibodies or with boiling citrate buffer, pH 6.0 for 20 min in a microwave oven for nuclear markers.
The following primary antibodies were applied overnight at 4°C: rabbit polyclonal anti-TGF-β (1:50 dilution; Abcam), rabbit polyclonal anti-fibronectin (1:100 dilution; Sigma), rabbit polyclonal anti-collagen type IV (1:100 dilution; BioTrend), and rabbit polyclonal anti-Runx2 (1:50 dilution; Abcam). Biotinylated secondary antibody (anti-rabbit, BioGenex) was applied for 20 min at 37°C, and then the slides were exposed to a streptavidin-biotin, alkaline phosphatase complex. Negative controls were performed by omitting the primary antibody.
Analysis was performed using a semiquantitative scoring system (0–12). The area which was positive rated (0–4; 0, none; 1, <10%; 2, 10–50%; 3, 51–75%; and 4, >75%) was multiplied by an index of intensity (0–3; 0, none; 1, weak; 2, moderate; and 3, strong). In the case of nuclear stainings, the number of positive cells per high-power visual field (×400) was determined.
Western blot analysis.
Aortic tissue was homogenized in 250 μl of homogenization buffer using an electronic stirrer. Protein concentration was determined by the Bradford method (4). Due to the limited amount of protein (300–400 μg) obtained from the mouse aorta, protocols were all standardized in a prior pilot study to evaluate drug dosage and toxicity.
Equal amounts (40 μg) of total protein were then separated by electrophoresis on polyacrylamide-SDS gels and transferred onto a polyvinylidene difluoride membrane (Immobilon-P membranes; Millipore, Bedford, MA) using the semidry method.
The membranes were washed and blocked with 5% nonfat dry milk in TBS-T buffer (pH 7.6) for 1 h. Thereafter, the membranes were incubated overnight at 4°C with the following primary antibodies: rabbit anti-VDR (Santa Cruz Biotechnology), rabbit anti-Runx2 (Abcam), goat anti-bone morphogenic protein-2 (BMP2; Santa Cruz Biotechnology), goat anti-BMP7 (Santa Cruz Biotechnology), mouse anti- MGP (Immundiagnostik), rabbit anti-osteoprotegerin (OPG; Santa Cruz Biotechnology), rabbit anti-Receptor Activator of NF-κB Ligand (RANKL; Santa Cruz Biotechnology), rabbit anti-TGF-β1 (Santa Cruz Biotechnology), goat anti-collagen type I (Santa Cruz Biotechnology), and mouse anti-β-actin (Santa Cruz Biotechnology) for the solvent. After washing, membranes were incubated with the horseradish peroxidase-conjugated secondary antibody (1:5,000, Santa Cruz Biotechnology, Heidelberg, Germany) and then visualized with an ECL kit (GE Healthcare, Buckinghamshire, UK) according to the manufacturer's instructions. Specific bands were quantified by densitometric analyses using image-analysis software (Image J 1.39u; http://rsb.info.nih.gov/ij/).
Results are presented as means ± SD. After testing for normal distribution, one-way ANOVA followed by Tukey's test was performed using Prism 5.0 software (GraphPad Software, San Diego, CA). A nonpaired t-test was performed when only two groups were compared. P < 0.05 was considered significant.
Biometric data and blood pressure.
Treatment with calcitriol caused a transient loss of body weight (at week 6, 25.1 ± 3.3 g UNX calcitriol vs. 28.6 ± 2.9 g UNX solvent, P < 0.01), but calcitriol-treated animals spontaneously regained body weight thereafter. By the end of the study there was no significant difference in body weight.
Blood pressure by the tail-cuff method was not significantly different among the groups (Table 1).
Creatinine clearance and albuminuria.
Although no significant difference in serum creatinine was detected between sham-operated and UNX animals (0.18 ± 0.01 mg/dl sham-operated vs. 0.21 ± 0.02 mg/dl UNX, P = 0.07), when analyzed together, sham-operated animals had a significantly higher creatinine clearance compared with UNX (220 ± 61 μl/min sham-operated vs. 165 ± 33 μl/min UNX, P = 0.01). When the individual groups were analyzed, however, no significant difference was seen. Albumin excretion was not significantly different among the groups. The results are presented in Table 1.
Neither UNX nor the treatments influenced the concentrations of total cholesterol, LDL cholesterol, or triglycerides (Table 1). HDL concentrations were also not significantly different between groups.
Ionized calcium, phosphorus.
A trend toward higher levels of ionized calcium was observed in UNX animals treated with calcitriol compared with solvent, but this was not statistically significant (P = 0.15). In contrast, the phosphorous concentration was significantly higher in the UNX calcitriol group (3.34 ± 0.66 mmol/l) compared with UNX solvent (2.67 ± 0.37 mmol/l, P < 0.05), but not to UNX paricalcitol (2.75 ± 0.26 mmol/l) (Fig. 1).
Morphology of the aorta.
With the exception of the UNX paricalcitol group, the aorta wall thickness was significantly higher in all UNX groups compared with sham-operated paricalcitol animals.
The wall-to-lumen ratio measured in glutaraldehyde-fixed semithin sections was significantly higher in UNX calcitriol than in UNX paricalcitol (P < 0.05) and the respective sham-operated groups (Table 2).
In all UNX groups with the exception of UNX paricalcitol, maximal plaque height was significantly higher than in the sham-operated solvent group. The intima/media ratio was increased in UNX solvent, UNX, and sham calcitriol, but not in UNX and sham paricalcitol animals (Table 2).
The increased intima/media ratio (UNX solvent and UNX as well as sham calcitriol) reflects mainly plaque enlargement in these groups (Table 2). Figure 2 contains representative samples of the alterations described above.
Aortic wall remodeling.
In the intima layer and plaques, the collagen content by Sirius red staining was not significantly different among the groups (Table 3 and Fig. 3). In the media, collagen expression was more pronounced in the UNX calcitriol, but nonsignificantly higher in UNX paricalcitol compared with the sham-operated solvent and sham-operated paricalcitol groups.
In the intima, immunohistochemical staining of plaques for collagen IV was not significantly different among the groups. In the media, collagen IV was markedly elevated in the UNX solvent compared with sham-operated and UNX paricalcitol groups, but not to UNX calcitriol (Table 4).
Collagen I expression by Western blotting was markedly increased in the aorta of UNX calcitriol compared with all other groups (Fig. 4). In UNX paricalcitol, there was no elevation of collagen I.
Expression of TGF-β1 by Western blotting was significantly higher in UNX solvent and UNX calcitriol compared with sham-operated solvent and sham-operated paricalcitol groups (see Table 6). In UNX paricalcitol, no significant elevation was observed (Fig. 4). By immunohistology, expression of TGF-β1 in the intima was significantly higher in UNX solvent and UNX calcitriol groups compared with UNX paricalcitol (Table 4). In the media, TGF-β1 expression was significantly higher in UNX solvent and UNX calcitriol groups, but not UNX paricalcitol compared with all sham-operated groups. The findings of TGF-β1 by immunohistochemistry are illustrated by Fig. 5.
Calcification of aortic plaques (intima calcification) was not seen in sham-operated animals treated with solvent, or in sham-operated animals treated with paricalcitol, but scattered plaque calcification was seen in plaques of 2 of 16 sham-operated animals treated with calcitriol. Similarly, in UNX animals calcification of aortic plaques was not observed in animals treated with solvent or paricalcitol. In contrast, signs of massive plaque calcification were seen in 6 of 16 of UNX calcitriol-treated animals.
Calcification of the aortic media (Mönckeberg sclerosis) was not seen in sham-operated animals or animals treated with solvent or paricalcitol. In contrast, striking media calcification was seen in 4 of 16 sham-operated calcitriol-treated animals. In UNX animals, calcification of the aortic media was absent in animals treated with solvent; scarce calcifications of the media were seen in 4 of 16 paricalcitol-treated animals, but signs of massive calcification were seen in most (9 of 16) of the calcitriol-treated UNX animals.
In the aortic intima of the UNX calcitriol, von Kossa staining was significantly more marked in UNX calcitriol compared with sham-operated and UNX solvent groups. In the media, in UNX calcitriol von Kossa staining was significantly more marked compared with all other groups with the exception of the sham-operated calcitriol (Table 3).
Expression of proteins related to calcification.
Runx-2-positive cells were seen in intima plaques as well as in the media, even in the absence of demonstrable calcifications (Fig. 6).
In sham-operated animals, Runx-2-positive cells in intima and media of mice treated with solvent or paricalcitol were sparse. The number was variable in calcitriol-treated animals. In the sham-operated calcitriol group, staining of Runx-2 was restricted to the media.
In UNX, the number of Runx-2-positive cells was low in animals treated with solvent and paricalcitol, but massively elevated in calcitriol-treated animals in both intima and media.
These immunohistochemical differences were even more pronounced by Western blotting (Table 5).
In animals treated with solvent, VDR expression was not significantly different between sham-operated and UNX. In contrast, VDR expression was upregulated in both paricalcitol- and calcitriol-treated UNX, but not in paricalcitol- and calcitriol-treated sham-operated animals (Table 5).
Only trace expression of RANKL was seen in sham-operated animals, with the exception of some expression in calcitriol-treated sham-operated animals. In UNX animals, trace expression was seen in solvent- and paricalcitol-treated animals, but expression was pronounced in calcitriol-treated animals.
The expression of osteoprotegerin was highly variable and not significantly different among the groups. Nevertheless, a significant difference of the RANKL/OPG ratio was seen between paricalcitol-treated sham-operated and calcitriol-treated (sham-operated and UNX) animals (Table 5). In UNX animals, no difference in the RANKL/OPG ratio between the treatment groups was observed. Representative blot samples are given in Fig. 4.
Expression of BMP 2 by Western blotting was significantly higher in UNX calcitriol, but not in UNX paricalcitol compared with sham-operated and UNX solvent groups (Table 6).
No difference in BMP 7 or MGP expression was seen among the groups.
The present study confirms previous observations (5, 6, 17) in ApoE−/− mice, an animal model with dyslipidemia and high oxidative stress, which had documented more pronounced atherosclerosis after 12 wk at the time of UNX and 24 wk at death, even in the absence of dietary interventions. Twelve weeks after UNX, a significant increase in the average diameters of the aortic wall, the wall-to-lumen ratio, and maximal plaque height was found.
Despite a minor transient difference in body weight between calcitriol- and paricalcitol-treated UNX animals, the final body weight was similar in the three UNX groups. Neither UNX nor any treatment significantly changed total cholesterol, triglycerides, HDL and LDL, which is in line with the clinical experience (33). A positive impact of calcitriol and paricalcitol on albumin excretion had been reported in SNX rats (28, 35), but in the present UNX model urinary albumin was not affected by UNX, calcitriol, or paricalcitol.
The main point in the present study is the difference between the reaction of the aorta of ApoE−/− mice in response to treatment with paricalcitol and calcitriol, respectively. This difference in plaque morphology and biochemistry is not explained by a difference in blood pressure, at least by tail plethysmography.
Calcitriol promoted intima hypertrophy and induced marked calcification of atherosclerotic lesions in parallel with concomitant Runx2-positive cells in some animals. In contrast, paricalcitol treatment prevented the elevation of TGF-β expression in plaque and collagen IV in the media and caused only spotty calcification. With respect to all structural parameters of the aorta, the values in UNX paricalcitol animals did not differ from those in sham-operated animals. All these parameters were significantly elevated in the UNX solvent group.
In paricalcitol- and calcitriol-treated UNX animals, serum calcium was not significantly different from solvent-treated UNX, although a tendency for higher values was noted in calcitriol-treated UNX. A significant difference in serum phosphate concentration was noted, however, between solvent-treated and calcitriol-treated UNX, but not paricalcitol-treated UNX animals. Whether this effect was due to increased intestinal phosphate absorption induced by calcitriol, changes in renal phosphate excretion, or phosphate mobilization from bone could not be established.
The two vitamin D compounds were dosed in a ratio 1:3.3 (calcitriol:paricalcitol) as proposed for the treatment of renal patients (31). The employed doses of both components and route of administration are in accordance to those currently used in experimental studies involving VDR activation (24, 25). Nevertheless, serum parathyroid hormone (PTH) was not measured in our study, and therefore a potential direct influence of PTH concentration in vascular remodeling by UNX ApoE−/− could not be evaluated. Moreover, a possible distinct inhibition of PTH in our model by the two compounds cannot be excluded.
The following significant differences in morphometric parameters of the aorta in favor of paricalcitol were noted between UNX mice treated with calcitriol and paricalcitol, respectively: wall-to-lumen ratio, intima/media thickness, and maximal plaque height. The number of plaques per unit area of the aorta was not different between UNX mice treated with paricalcitol and calcitriol, respectively. Since the number of plaques in the aorta was not different between sham-operated and UNX mice, any observed differences must be due to a different rate of expansion of preexisting plaques. Nevertheless, no sign of cell proliferation in plaques was observed, as staining for proliferating cell nuclear antigen was negative (data not shown).
Even in sham-operated animals, administration of calcitriol caused a significant increase in intima thickness (increase in intima/media ratio). This was not seen with paricalcitol. The finding is in line with the observation of Lamawansa et al. (20), who reported intima hyperplasia in calcitriol-treated rats in an experimental model of arterial balloon injury.
The composition of aortic plaques was significantly different between calcitriol- and paricalcitol-treated UNX animals: calcification was increased after UNX in calcitriol-treated, but not in paricalcitol-treated mice. This was paralleled by increased expression of TGF-β1 in the calcitriol-treated animals. Such an increase was even seen in the aortic plaques of UNX mice treated with calcitriol compared with the UNX mice treated with paricalcitol.
The inhibitory effect of VDR activation on TGF-β expression has already been documented in the kidney from studies evaluating nephropathy progression (24, 27, 32). In the aorta, the interpretation of the TGF-β1 findings is challenging: on the one hand, TGF-β expression is reduced by active vitamin D (22, 24, 32), and on the other hand TGF-β is upregulated by calcification (18). In our study, we found upregulation of TGF-β1 by UNX, which was abrogated by paricalcitol, but not by calcitriol.
We believe that the elevation of TGF-β1 expression in the UNX solvent and UNX calcitriol groups reflects a dynamic injury situation secondary to the UNX model or to vascular calcification with concomitant vascular TGF-β-mediated repair. This repair led to an excessive deposit of collagen IV by UNX solvent or both collagens I and IV by UNX calcitriol and to an increase in the wall-to-lumen ratio in these groups. In the specific case of UNX calcitriol, the anti-inflammatory actions of VDR activation were possibly overcome by the profibrotic effect of vascular calcifications observed in these animals.
In UNX calcitriol-treated mice, both calcification of plaques (as seen in atherosclerosis) and of media (as seen in Moenckeberg sclerosis) were noted in the aorta. This difference was paralleled by the expression of Runx2, a marker pointing to an osteoblastic phenotype.
Cell culture studies had shown that high phosphate triggers vascular calcification (19) and that serum of uremic patients triggers calcification even in the presence of normal phosphorous concentrations (25). Incubation of rat vascular smooth muscle cells with paricacitol and calcitriol showed that calcium incorporation curves of both agents are distinct irrespective of Ca × P concentrations, which is explained by a higher RANKL/OPG ratio induced by calcitriol (7). Nevertheless, despite not being significant, the tendency for higher phosphate concentrations in calcitriol- compared with paricalcitol-treated UNX Apo E−/− mice from our study is therefore of interest and may also be responsible for the occurrence of vascular calcifications and the higher synthesis of collagen I in calcitriol-treated animals (14, 19).
Calcification is the result of the balance between promoters and inhibitors. It is of interest that expression of calcification inhibitors (BMP7, OPG, MGP) was not significantly altered by UNX in the present model. In contrast, expression of calcification promoters (RUNX2, BMP2, RANKL) was significantly higher in calcitriol- compared with paricalcitol-treated animals. Whether the difference between paricalcitol and calcitriol treatment in UNX animals is driven by the difference in serum phosphate or PTH concentration, by intrinsic differences between the two compounds, or by other factor is unknown.
The results of the present study in UNX ApoE−/− mice support the hypothesis that selective VDR activation by paricalcitol (compared with calcitriol) results in less plaque growth and calcification. The extent to which this model can be extrapolated to humans is unclear.
L. E. Becker was supported by a CAPES/DAAD-funded fellowship. This work was supported by Abbott Laboratories, Chicago, IL.
No conflicts of interest, financial or otherwise, are declared by the authors.
The authors thank Heike Ziebart, Zlata Antoni, Annett Müller, Monika Weckbach, Peter Rieger, Josef Strunk, and Annemarie Wiss for outstanding technical help.
Present address of A. Geldyyev: Div. of Pathology, Turkmen State Medical Institute, Ashgabat, Turkmenistan.
Present address of G. Kökeny: Dept. of Pathophysiology, Semmelweis University, Budapest, Hungary.
- Copyright © 2011 the American Physiological Society